Current research trends and application areas of

Advanced Engineering Informatics 33 (2017) 112–131

Contents lists available at ScienceDirect

Advanced Engineering Informatics journal homepage: www.elsevier.com/locate/aei

Full length article

Current research trends and application areas of fuzzy and hybrid methods to the risk assessment of construction projects Muhammad Saiful Islam a, Madhav Prasad Nepal a,⇑, Martin Skitmore a,b, Meghdad Attarzadeh c a

School of Civil Engineering and Built Environment, Science and Engineering Faculty, Queensland University of Technology (QUT), 2 George Street, Brisbane, QLD 4000, Australia Research Institute of Complex Engineering and Management, School of Economics and Management, Tongji University, Shanghai, China c School of Civil and Environmental Engineering, Nanyang Technological University, N1-01a-29, 50 Nanyang Avenue, 639798, Singapore b

a r t i c l e

i n f o

Article history: Received 18 January 2017 Received in revised form 10 May 2017 Accepted 1 June 2017

Keywords: Construction projects Risk assessment Fuzzy logic Bayesian belief network Decision making

a b s t r a c t Fuzzy and hybrid methods have been increasingly used in construction risk management research and this study aims to compile and analyse the basic concepts and methods applied in this field to date. A content analysis is made of a comprehensive literature review of publications during 2005–2017. It is found that the nature of complex projects is such that most risks are interdependent of each other. Therefore, a fuzzy structured method such as the fuzzy analytical network process (FANP) has frequently been used for different complex projects. However, the application of FANP is limited because of the tedious and lengthy calculations required for the pairwise comparisons needed and an inability to incorporate new information into the risk structure. To overcome this constraint, a fuzzy Bayesian belief network (FBBN) has been increasingly used for risk assessment. Further project-specific studies based on FBBN are recommended to justify its broader application. Beyond fuzzy methods, the Credal network – an extended form of Bayesian network- is found to have potential for risk assessment under uncertainty. Ó 2017 Published by Elsevier Ltd.

0. List of acronyms

AC ACME AEIC AHP AMM ANFIS ANN ANP ASCE ASME BBN BOT CAJCIE

Automation in Construction Archives of Civil and Mechanical Engineering Architectural Engineering Institute Conference Analytical hierarchy process Applied Mathematical Modelling Adaptive neuro-fuzzy inference system Artificial neural networks Analytical network process American Society of Civil Engineers American Society of Mechanical Engineers Bayesian belief networks Build-operate-transfer Computer-Aided Journal of Civil and Infrastructure Engineering

⇑ Corresponding author. E-mail addresses: [email protected] (M.S. Islam), madhav. [email protected] (M.P. Nepal), [email protected] (M. Skitmore), [email protected] (M. Attarzadeh). http://dx.doi.org/10.1016/j.aei.2017.06.001 1474-0346/Ó 2017 Published by Elsevier Ltd.

CBR CEM CFPR CIE CJCE CME COPRAS DAG EJGE EI ER ESA ETA FAHP F-ANN FANP F-BBN FCE FCOPRAS FES FGDM FLIMAP

Case-based reasoning Construction and Engineering Management Consistent fuzzy preference relations Computer and Industrial Engineering Canadian Journal of Civil Engineering Construction Management and Economics Complex proportional assessment Directed acyclic graph Electronic Journal of Geotechnical Engineering Ekonomska Istrazivanja Evidential reasoning Expert Systems with Application Event tree analysis Fuzzy analytical hierarchy process Fuzzy artificial neural network Fuzzy analytical network process Fuzzy Bayesian belief network Fuzzy comprehensive evaluation Fuzzy complex proportional assessment Fuzzy expert system Fuzzy group decision making Fuzzy linear programming technique for multidimensional analysis of preference

M.S. Islam et al. / Advanced Engineering Informatics 33 (2017) 112–131

FMADR FMCS FMEA FR-MCS FST FTA GA HFES IEEE ICTE

IJET IJPM JAS JCEM JCCE JCivEM JCP JLPPI JOMAE JPSEP JRUES JVC KSCE JCE LINMAP MCDA MCDM MCS PCS PPP PSBS RA RMTTs RPN SEP SIE SJR SS SVM TUST TOPSIS UBC UHV VIKOR

Fuzzy Multi Attribute Direct Rating Fuzzy Monte Carlo simulation Failure mode and effect analysis Fuzzy randomness Monte Carlo simulation Fuzzy set theory Fault tree analysis Genetic algorithm Hierarchical fuzzy expert system Institute of Electrical and Electronics Engineers International Conference on Transportation Engineering IFEMCDM Integrated Fuzzy Entropyweight Multiple Criteria Decision Making International Journal of Engineering and Technology International Journal of Project Management Journal of Applied Sciences Journal of Construction Engineering and Management Journal of Computing in Civil Engineering Journal of Civil Engineering and Management Journal of Cleaner Production Journal of Loss Prevention in the Process Industries Journal of Offshore Mechanics and Arctic Engineering Journal of Pipeline Systems Engineering and Practice Journal of Risk and Uncertainty in Engineering Systems, Part A: Civil Engineering Joint venture contract KSCE Journal of Civil Engineering Linear programming technique for multidimensional analysis of preference Multiple criteria decision analysis Multi-criteria decision making Monte Carlo simulation Procedia Computer Science Public private partnership Procedia Social and Behavioral Sciences Risk Analysis Risk management tools and techniques Risk priority number System Engineering Procedia Structure and Infrastructure Engineering Scimago Journal & Country Rank Safety Science Support vector machine Tunneling and Underground Space Technology Techniques for order of preference by similarity to an ideal solution The University of British Columbia Ultra high voltage Vise Kriterijumska Optimizacija I Kompromisno Resenje (Multicriteria Optimization and Compromise Solution)

113

Failure to make a timely assessment of risks and their impact on project objectives (e.g., project duration and cost) can hinder project success [7]. Construction risks are complicated, uncertain and subjective by nature due to the unique features of project-based activities [8]. Historical data of similar projects do not always represent the risk status of new projects, which leads to a dependence on expert judgment. Expert judgments are usually uncertain and subjective due to a vague and imprecise understanding of project risks. Many advanced methods exist for assessing the risks of construction projects. These can be broadly categorised into three types, i.e., indexing, matrix and probabilistic methods [3,9]. Indexing methods are the most popular due to their simplicity of application based on expert judgments, but cannot provide accurate results for complex projects, where the risks involve uncertainties. Matrix methods are applicable for analysing expert judgments and providing a better result for complex projects, but are also incapable of capturing subjectivity and uncertainty in the data. Probabilistic methods provide a robust process for risk assessment of complex projects but need a large amount of data from similar, previously constructed, projects. Thus, probabilistic methods are not appropriate for assessing the construction risks of complex projects because of imprecise and insufficient data [10]. In contrast, fuzzy methods are very efficient in modelling the uncertainties encountered in expert judgments and have therefore been frequently and widely used as independent or hybridised methods of construction risk assessment for the last two decades [11]. There has been an increasing number of publications concerning construction project risk-management in recent years using fuzzy logic and hybrid methods. This study responds to the need for a better understanding of the potential applications of fuzzy and hybrid methods for construction risk management. Chan et al. [11] summarised and critiqued research into ‘‘fuzzy techniques” in construction management published between 1996 and 2005. They presented a thorough review of the application of fuzzy logic, fuzzy set theory and hybrid fuzzy methods. During this period, fuzzy logic was hybridised mainly by artificial neural networks (ANN), and the genetic algorithm [11]. Recent studies and research trends have revealed the popularity of using multicriteria decision making (MCDM) methods to hybridise fuzzy techniques for risk assessment [12]. Rezakhani [6] presented another good review of fuzzy logic for project risk management. However, the study did not cluster fuzzy and hybrid methods, and no project specific applications of these methods were provided. There are many risk assessment methods available for a specific project or projects of a similar type. It is within this context and in the context of construction project risk management that this paper aims to provide a detailed review of research into fuzzy and hybrid methods to delineate their application areas, identify research gaps and guide potential research directions. The remainder of this paper is devoted to the research methodology, construction riskmanagement tools and techniques, fuzzy-based construction risk management methods, applications of fuzzy and hybrid methods, discussion, future research directions and conclusions.

1. Introduction Project risk can be defined as an uncertain event that leads to failing to achieve at least one project objective [1,2]. The risk management process can improve project performance by controlling the consequences of risky events on project objectives [3]. It is recognised that it is possible to manage risks but not eradicate them [4]. Risk management involves several steps, such as risk identification, analysis, assessment, prioritisation and responding to project risks with the aim of enhancing opportunities and reducing negative consequences [5,6]. Of these, risk assessment is an important aid in decision science for managing uncertain events.

2. Research methodology This study is based on a comprehensive literature review of recently published (2005–2017) relevant papers. The literature is drawn from the top quality journals in the field of construction engineering and project management listed in the Scimago Journal & Country Rank (SJR) list. Three additional relevant papers from three journals not listed in the SJR are also included in the review. The most frequently cited journals in this study are: (1) Expert Systems with Application, Elsevier; (2) the Journal of Construction

114

M.S. Islam et al. / Advanced Engineering Informatics 33 (2017) 112–131

Engineering and Management, ASCE; (3) the International Journal of Project Management, Elsevier; (4) Construction Management and Economics, Taylor & Francis Group; (5) Automation in Construction, Elsevier; (6) Risk Analysis, Wiley-Blackwell; (7) IEEE Transactions on Engineering Management, IEEE; (8) the Journal of Civil Engineering and Management; (9) the Canadian Journal of Civil Engineering, CSCE; (10) the Journal of Computing in Civil Engineering, ASCE; (11) the ASCE-ASME Journal of Risk and Uncertainty in Engineering Systems, Part A: Civil Engineering; (12) the ASCE-ASME Journal of Risk and Uncertainty in Engineering Systems, Part B: Mechanical Engineering; (13) Safety Science; (14) the Computer-Aided Journal of Civil and Infrastructure Engineering; and (15) Fuzzy Sets and Systems, Elsevier. Four books and some conference proceedings published by IEEE, ASCE and Elsevier were also used. A keyword search for ‘‘fuzzy logic”, ‘‘fuzzy set”, ‘‘fuzzy risk assessment”, ‘‘risk vulnerability assessment”, ‘‘uncertainty assessment” and ‘‘construction project”, which are frequently used in fuzzy based risk management papers [11], was conducted initially to retrieve fuzzy papers from the ‘‘Google Scholar”, ‘‘Science Direct”, ‘‘Web of Science”, ‘‘ASCE Library”, Wiley Online Library and ‘‘University Online Library” search engines. The keywords ‘‘vulnerability”, ‘‘reliability” and ‘‘probability” were further used to widen the search. Altogether, 133 research papers containing the term ‘‘fuzzy” in their titles, abstract or keywords were identified. Of these, 39 papers were outside the construction project-management domain, such as health and safety, telecommunication, industrial manufacturing. Table 1 lists the remaining 94 papers. Of these, 11 papers did not involve any specific application of fuzzy and hybrid methods to construction risk management. This left 83 papers, as shown in Tables 2–4, being available for detailed analysis. An additional 76 papers concerning fuzzy and beyond fuzzy-hybrid based methods were also included. There are essentially two types of analyses for literature-based independent studies — meta-analysis and content analysis [11,13]. Content analysis (i.e. paper title, abstract, keywords, methodologies, and model demonstrations, or application areas) was employed to summarise the basic concepts of the fuzzy-based methods/models, discuss the project specific methods/models used in risk assessment, provide a critique of their applications and pinpoint specific research gaps. The fuzzy-based methods are classified into basic fuzzy, extended fuzzy and hybrid fuzzy methods [9]. The application areas of fuzzy-based risk management are classified into six groups: building construction, roads and highways, subways/tunnels, pipelines, power generation and transmission, and other or overall (not specified) construction projects. Of the papers analysed, 11 papers are related to building projects, 2 to real estate, 26 to road and highway projects (Table 2), 14 to subways/tunnels and 6 to pipeline projects (Table 3), 8 to the power sector, 16 to others (i.e. rigging pipes and welding, sanitary tanks, and, oil and gas reservoirs) or overall (not specified) construction projects (Table 4). Table 1 also summarises the current state of knowledge of risk management for construction projects based on fuzzy and hybridised tools and techniques.

basic fuzzy method can be defined as representing the basic concept of fuzzy logic and fuzzy set theory. The extended fuzzy method has modified algorithms based on fuzzy theory but not modified by other independent methods such as fuzzy arithmetic, fuzzy synthetic evaluation, fuzzy expert system, fuzzy Mamdani inference, fuzzy comprehensive evaluation and fuzzy consensus qualitative analysis. The hybrid fuzzy method represents a combination of fuzzy and other independent methods. It involves different types, such as fuzzy probability methods, fuzzy matrix methods, fuzzy structured methods, the fuzzy cloud model and fuzzy integral process. The fuzzy probability methods comprise the event tree, fault tree, Monte Carlo simulation (MCS), Bayesian probability theory, artificial neural network (ANN), and failure mode and effect analysis (FMEA). Fuzzy matrix methods are techniques for ordering preferences by the similarity to an ideal solution (TOPSIS), Vise Kriterijumska Optimizacija I Kompromisno Resenje (Multicriteria Optimization and Compromise Solution or VIKOR) and complex proportional assessment (COPRAS). Fuzzy structured methods comprise the analytical hierarchy process (AHP), analytical network process (ANP) and Bayesian belief networks (BBN). The following subsections briefly discuss fuzzy and hybrid methods most frequently in use. 3.1. Basic fuzzy methods 3.1.1. Fuzzy logic Construction risks are still managed based on expert judgments and experience [103] and hence the data for risk studies is mostly qualitative. Fuzzy logic has been used in risk evaluation for long time, as it can be used to develop models on the basis of both qualitative data and quantitative values from historical records [58], and is therefore a very effective management technique for achieving the objectives of construction projects under uncertainties, impression and biasness [29]. Lyons and Skitmore [103] and Novak [104] described the common features of fuzzy logic as involving the following basic steps: (1) define and measure the likelihood of occurrence and severity of the risk in terms of verbal opinions and transform them into fuzzy numbers accordingly; (2) define a fuzzy inference to make a network between input and output parameters using fuzzy IF-THEN rules and/or ‘‘fuzzy arithmetic operators”; and (3) defuzzify the fuzzy outcomes into numerical values using appropriate quantifiers. Although existing fuzzy logic is a well-established theory, the lack of appropriate techniques to address fuzzy consistency and fuzzy priority vectors, together with the complicated operations involved, undermines its practical application [55]. Fuzzy logic is in particular need of improvement for modelling qualitative data elicited from expert opinion using natural language. If there is a lack of knowledge about project risks, the qualitative data elicited from experts contains vagueness and uncertainty [105]. Novak [104] discussed the potential of mathematical fuzzy logic to improve models affected by vagueness and suggested combining fuzzy logic with probability theory to capture the uncertainties involved. Further improvement of fuzzy logic is also necessary for modelling evaluative appropriate linguistic expressions of risk and developing suitable aggregation rules to quantify the linguistic expressions for risk ranking [104,106].

3. Fuzzy based construction risk management methods Numerous methods have already been applied with the aim of developing construction risk management models (Table 1). Construction project risks are uncertain and vague in nature, which has led to the applications of the fuzzy concept [8]. However, several drawbacks of fuzzy methods have been encountered and hybrid methods are increasingly being used [102]. Fuzzy-based methods can be classified into three broad groups of (1) basic fuzzy, (2) extended fuzzy and (3) hybrid fuzzy methods [11]. The

3.1.2. Fuzzy set theory Construction industry decision makers need to manage complex, dynamic problems under conditions of uncertainty [107]. Fuzzy set theory (FST) is a well-recognised decision support tool that allows the uncertainty of the events in risk assessment based on the experts’ linguistic evaluation approach to be handled [22]. This is an extended form of classical binary logic, where a problem is considered solely as having full or non-membership (i.e., 0 or 1). In practice, construction risks are not usually possible to define in

115

M.S. Islam et al. / Advanced Engineering Informatics 33 (2017) 112–131 Table 1 Summary of studies of risk management tools and techniques (RMTTs) for construction and infrastructure projects. Method/tool/theory

Keynotes

Author/s

Model and application

Reference

Fuzzy logic

Uncertainty, lack of information, experts’ knowledge. Multi-criteria decision analysis.

Jamshidi et al. (2013) Kucukali (2011) Ismail et al. (2008) Choi and Mahadevan (2008) Elbarkouky et al. (2015) Mao and Wu (2011) Pawan and Lorterapong (2015) Maravas and Pantouvakis (2012) Malek et al. (2015) Zho et al. (2013) Salah (2015)

Risk assessment and decision making model for assessing the level of risks in pipelines. Multi-criteria fuzzy rating tool. Risk assessment method for river-type hydropower plant projects. Fuzzy logic for identifying and ranking risk sources and factors in construction projects. Risk assessment model of a cable-stayed bridge by combining existing data and information with a systematic updating methodology. Fuzzy logic-based contingency determination tool.

[14]

FST

Fuzzy arithmetic Fuzzy real option FST

Significance of risk sources, expert interviews, cost effective method. Decision support system.

Uncertainties. Contingency determination. Fuzzy weighted average, reliability of value evaluation. Impression, vagueness, expert judgement.

Uncertainty, imprecision, feasibility studies.

Uncertainties, decision making.

FST and fuzzy probability theory FST, entropy weight method FST and Consensus aggregation, Delphi technique FST, evidential reasoning FGDM Fuzzy synthetic model Fuzzy synthetic evaluation approach

Enterprise risk management (ERM), maturity criteria, resource prioritisation. Qualitative and quantitative risk evaluation process, risk mapping. Uncertainties, multiple criteria for decision making. Linguistic framework, experts’ opinion aggregation. Information synthesising, decision support framework Risk assessment. Multiple risk factors, rapid assessment. Incomplete data. Vague environment.

Mokhtari et al. (2012) Wang and Elhag (2007) Abdul-Rahman et al. (2013)

Risk management attitude. Contractor’s risk analysis capability. Risk criticality analysis. Risk evaluation, cost variability, contract cost and final accounts.

Mu et al. (2014) Zhao et al. (2015) Ameyaw et al. (2015) Xu et al. (2010a) Xu et al. (2010b) Salawu and Abdullah (2015) Idrus et al. (2011) Fayek and Oduba (2005) Fares and Zayed (2010) Aboshady et al. (2013) Dikmen et al. (2007) Zhao and Li (2015)

Decision making based on expert knowledge. Critical risk group, project risk level. Risk management maturity, decision support system FES

HFES FES, fault tree, even tree Fuzzy set and Influence diagram Fuzzy comprehensive evaluation (FCE) and cloud method FCE

Fuzzy logic, least square, and support vector machine Fuzzy arithmetic and fault-tree

Ji et al. (2015)

Subjective judgement. Risk analysis. Industrial construction. Linguistic prediction. Failure risk analysis.

[15] [16] [17]

[18]

Risk analysis for real estate investment.

[19]

An integrated framework for risk assessment and time contingency modelling.

[20]

Cash flow calculation model for projects with fuzzy durations and/or costs, with application to a road construction project.

[21]

Fuzzy logic-based risk assessment tool for determining the extent of concrete deterioration. Model for assessment of the ERM maturity level of construction firms. Risk management model for risk assessment, ownership, contingency determination and mitigation strategies. Integrating fuzzy set theory, the entropy weight method and multiple criteria decision-making method for risk assessment of hydropower stations. Fuzzy consensus qualitative risk analysis framework to identify and prioritise risks encountered in real estate projects.

[22] [23] [24] [25]

[26]

A framework to facilitate decision making under uncertainties in a sea port and terminals. Risk assessment model for bridge construction projects.

[27]

A fuzzy synthetic model to estimate construction risks especially for situations with incomplete data and vague environments. Contractor risk management capability (RMC) assessment model for subway projects in mainland China. Fuzzy synthetic evaluation-based risk assessment model for green building projects in Singapore. A fuzzy model for the assessment of cost variation risks in public construction projects. A fuzzy risk allocation model for determining equitable risk allocation in China PPP projects. A fuzzy synthetic evaluation model for risk assessment of China PPP projects. Approach for assessing the risk management capability of contractors in highway rehabilitation projects.

[29]

[28]

[30] [31] [32] [33] [34] [35]

Cost contingency estimation model for Malaysian building and infrastructure projects. Modelling industrial construction (rigging pipes and welding) productivity. Water main failure risk assessment.

[37]

A risk management framework for real estate projects.

[38]

Risk assessment model for rating cost overrun risks of international construction projects. Risk evaluation model for ultra-high voltage (UHV) power transmission projects.

[39]

Yu et al. (2015)

Risk evaluation of subway construction.

[41]

Risk assessment model for underground oil and gas reservoirs.

[42]

Artificial intelligence approach, differential evolution.

Wang et al. (2016) Cheng and Hoang (2014)

Risk scoring model for prioritising bridge maintenance works

[43]

Risk event analysis. Linguistics probability.

Abdelgawad and Fayek

Fuzzy fault tree analysis method for the assessment of risk events of pipe line projects.

[44]

Critical root causes, risk mitigation cost. Cost overrun risk. Randomness and discreteness. Uncertainty and vagueness. Risk weight, expert weight, and expert scoring method. Risk likelihood and consequences.

[36]

[9]

[40]

(continued on next page)

116

M.S. Islam et al. / Advanced Engineering Informatics 33 (2017) 112–131

Table 1 (continued) Method/tool/theory

Keynotes

Author/s

Fuzzy sets, fault tree, Delphi technique Fuzzy logic, system dynamics FST, evidence theory, fault-tree and event-tree AHP

Privatised infrastructure, absent of past data, project specific risks. Complex structure, and dynamics in the nature of risks and uncertainties. Qualitative risk assessment, interdependencies between events.

(2011) Thomas et al. (2006) Nasirzadeh et al. (2008) Ferdous et al. (2011)

Fuzzy logic, FMEA

Cost estimation, uncertainties, responding to cost overrun risks. Identifying failure modes, risk priority number (RPN).

Fuzzy logic, AHP and FMEA

Relative measures of project complexity.

Fuzzy FMEA and FAHP

Mapping the relationship between impact, probability of occurrence, and detection.

Fuzzy, FMEA, Fault tree, Event tree

Lack of sufficient data, imprecise probability knowledge of risks.

FAHP

Risk assessment, linguistic variables, vagueness of expert judgment, pair-wise comparison, dynamic nature of the project.

Fuzzy FMEA

Subjective judgment, uncertainties. High complexity, risk criticality.

Fuzzy set, FMEA, VIKOR Fuzzy VIKOR

Failure mode detection, prioritisation.

Fuzzy logic, FMEA

Vidal et al. (2011) Jung et al. (2015) Mohammadi and Tavakolan (2013) Abdelgawad and Fayek (2010) Abdelgawad and Fayek (2012) Nieto-Morote and Ruz-Vila (2011) Nguyen et al. (2015) Zou and Li (2010) Li and Zou (2011) Zeng et al. (2007) Sharma et al. (2005) Cheng and Lu (2015) Chen and Wang (2009) Subramanyan et al. (2012) Zhang and Zou (2007) Abdollahzadeh and Rastgoo (2015) Sonmez (2011)

Fuzzy fault-tree and event-tree

Risk assessment. Likely cost of risk event

Neural network, bootstrap method, Bayesian probability CBR hybrid model

Relationship between risk factors and cost. Cost range estimation.

Predicting project duration and cost.

Koo et al. (2010)

Fuzzy logic, genetic algorithm, and neural networks ANFIS

Feasibility study, cost estimation at early stage under uncertainties.

Cheng et al. (2009)

Subjective judgement, non-linear relationships between risks. Allocating risk.

Wang and Elhag (2008) Imbeah and Guikema (2009) Farajian and Cui (2011) Khodakarami and Abdi (2014) Tesfamariam (2013) Khakzad et al. (2013) Medina-Oliva et al. (2009) Weber et al. (2012) Cárdenas et al. (2014)

Bayesian probability technique and utility theory

BBNs

Cost analysis, risk networking

Quantifying potential risks. Safety risk analysis. Safety, reliability and maintenance. Overview of Bayesian belief networks. BBN and sensitivity analysis

Risk networking, probability of possible failure events

Model and application

Reference

Risk assessment model for BOT-road projects.

[45]

Construction risk assessment model to capture the dynamics of risks and uncertainties. A qualitative risk assessment framework for managing uncertainties and interdependencies of risk events.

[46]

Defining and measuring relative project complexity using the analytic hierarchy process. Cost contingency modelling for identification and analysis of cost overrun risks of public construction projects. A risk assessment model by combining fuzzy logic, failure mode and effect analysis (FMEA) for a subway construction project. Assessing the level of criticality of risk events in the construction domain by combining fuzzy logic with FMEA and AHP. Comprehensive risk management framework to assess the expected cost of risk events in a pipeline project.

[48]

A hierarchical weighting method-based risk assessment method for construction projects.

[2]

Project complexity quantification process of transportation projects. Risk checklist development and risk assessment for subway projects. Risk assessment model for a PPP expressway project.

[7]

[54]

Risk assessment method for a building project.

[55]

Risk assessment model for system safety and reliability analysis. Construction risk assessment method for pipe-jacking project.

[56]

[47]

[49] [50]

[51]

[52]

[53]

[10]

Risk assessment of international projects using a 3-level hierarchy structure. Expert opinion-based risk assessment model applied.

[57]

Risk assessment model for China joint-venture projects.

[59]

Risk assessment of bridge construction projects.

[3]

A model for cost range estimation of building projects.

[60]

A CBR-based model by integrating ANN, genetic algorithm and MCS for predicting project cost and duration of multi-family housing projects. Evolutionary Fuzzy Neural Inference Model for conceptual cost estimation of construction projects.

[61]

ANFIS-based risk assessment model for bridge maintenance projects. The application of Advanced Programmatic Risk Analysis and Management Model (APRAM) for managing schedule, cost and quality risks in the construction industry. Multi-objective decision support system for PPP funding decisions and portfolio analysis. Modelling dependencies between cost items using Bayesian networks.

[63]

[58]

[62]

[64]

[65] [66]

Risk assessment of bridges.

[67]

Bayesian network approach for quantitative risk analysis of offshore drilling operations. A review of the application of Bayesian networks.

[68]

A review of Bayesian belief networks for modelling dependability, risk analysis and maintenance problems. Representing and analysing risk-related knowledge using Bayesian networks for risk mitigation and resource allocation in tunnel construction.

[69] [70] [71]

117

M.S. Islam et al. / Advanced Engineering Informatics 33 (2017) 112–131 Table 1 (continued) Method/tool/theory

Keynotes

Author/s

Model and application

Reference

Fuzzy- Bayesian network approach

Safety risk analysis. Causal relationships between organisational factors.

Ren et al. (2009) Li et al. (2012)

[72]

Safety risk analysis, capturing uncertainties.

Zhang et al. (2016) Kabir et al. (2016) Lin and Jianping (2011) Ebrahimnejad et al. (2012) Ebrat and Ghodsi (2014) Taylan et al. (2014) Tamosaitience et al. (2013)

Modelling causal relationships between the factors causing accidents in offshore operations. Modelling human reliability factors to find the root causes of human error in nuclear power plants. Modelling tunnel-induced safety risks for pipeline damage using fuzzy Bayesian network-based approach. Modelling safety risks for identifying the critical causes of oil and gas pipelines failure. Risk assessment model for identifying critical risks in the construction of a university campus project. Two-phase fuzzy group decision-making approach to facilitate the project selection process. An adaptive neuro-fuzzy inference system for risk assessment of construction projects. Project risk evaluation and selection with incomplete and uncertain information. Integrated risk assessment model using fuzzy AHP and fuzzy TOPSIS methods to assess the overall risks of construction projects. A fuzzy TOPSIS method to aid the selection of a suitable risk assessment model in the construction industry. Fuzzy logic-based multi-criteria decision making for tunnelling projects. Risk identification and ranking for power plant projects using a fuzzy-ANP and fuzzy TOPSIS method. Risk evaluation method for construction projects in Iran.

Safety assessment. Fuzzy comprehensive evaluation, ANP Fuzzy set, ANP, VIKOR

High-risk projects, risk hierarchy, risk networks. Group decision making, risk and uncertainty.

ANFIS

Risk identification and priorities in different project phases. Project planning. Construction project selection, group decision making in a fuzzy environment. Multi-criteria decision-making.

FAHP, and fuzzyTOPSIS Fuzzy TOPSIS

Hybrid neural network; conceptual cost estimation, risk allocation.

Karimiazari et al. (2011) Fouladgar et al. (2012) Zegordi et al. (2012) Ebrahimnejad et al. (2008) Ebrahimnejad et al. (2010) Yazdani et al. (2011) Kuo and Lu (2013) Duran et al. (2009) Cheng et al. (2010) Jin (2010)

Fuzzy logic, ANN, fast messy GA, component ratios method ANP

Cost estimation, uncertainties and lack of data.

Hsiao et al. (2012)

Decision support tool, risk relationship.

FANP

Multi-criteria decision analysis.

Bu-Qammaz et al. (2009) Shafiee (2015)

FANP, fuzzy-TOPSIS Fuzzy with TOPSIS and LIMAP Fuzzy-TOPSIS, and Fuzzy-LINMAP Fuzzy COPRAS CFPR, FMADR ANN F-ANN

Best alternative selection, uncertainties, imprecise data. Underground construction, risk evaluation, uncertainties. Hierarchical structure, dependencies among risks. Multi-attribute decision-making. Project complexities, uncertainties, numerous risks. Multi-criteria decision making, criterion weight, importance of alternatives. Uncertainties, insufficient information, subjective judgment. Cost estimation under uncertainties.

Risk prioritisation.

Fuzzy MCS FST, MCS Fuzzy randomness (FR), fuzzy MCS

Fuzzy discrete event simulation

Expert investigation, relation matrix, construction safety. Cost uncertainties, multi-criteria evaluation. Cost range estimation, extracting from experts. Risk and uncertainty modelling, combined propagation and analysis. Fuzzy logic.

Subjective uncertainty. Optimising construction resources.

Valipour et al. (2015) Yan et al. (2015) Wang et al. (2007) Shaheen et al. (2007) Möller and Beer (2008) Sadeghi et al. (2010) Heravi and Faeghi (2014) Attarzadeh et al. (2017) Sadeghi et al. (2015)

this way because of the complexity and uncertainty of the problems [11]. In contrast with binary logic, fuzziness is essential for the gradual transition of an element in a set from its membership to non-membership state [108]. Therefore, FST modifies basic binary logic to capture uncertainty and vagueness in defining risk. For example, risk is easy to define in such linguistic terms as ‘‘extremely high”, ‘‘very high”, ‘‘medium”, ‘‘low”, ‘‘very low” or ‘‘none”

[73] [74] [75] [76] [77] [78] [8] [79]

[80] [81] [82] [83]

Risk identification, assessment and ranking for BOT power plant projects Risks ranking and evaluation model for critical infrastructures.

[84]

Risk assessment approach under multi-criteria decisionmaking for metropolitan construction projects. Cost estimation model for shell and tube heat exchanger projects. Improving cost estimation precision.

[86]

F-ANN model to facilitate a decision support system for PPP projects. Cost estimation model for semiconductor hook-up construction projects.

[89]

A tool for estimating the level of risk of international projects.

[91]

Selection of a suitable risk mitigation strategy under uncertainty for offshore wind farms. A method of finding significant risks in freeway PPP projects.

[92] [93]

Risk analysis model for highway mountain tunnel.

[94]

Bid price evaluation model under multi-criteria decisionmaking. Fuzzy cost range estimation model for a sanitary tank project.

[95]

Non-traditional uncertainty models for engineering computation. Fuzzy Monte Carlo Simulation (FMCS) framework for risk analysis of construction projects. FMCS model for optimising time, cost and quality for dam construction. Fuzzy randomness-MCS model for evaluating risks and uncertainty of infrastructure projects. Event-based simulation model for the analysis of queues in asphalt paving operations.

[97]

[85]

[87] [88]

[90]

[96]

[98] [99] [100] [101]

instead of ‘‘risk” or ‘‘no risk”. Following this concept, Kangari [107] used fuzzy sets for the linguistic expression of risks (i.e. low, medium, high) and to elicit the risk factors from experts. Kangari and Riggs [109] developed a FST-based qualitative risk analysis model for solving the ill-defined nature of risk, comprising a three-step risk analysis system of natural representation by FST, fuzzy set evaluation of risks, and linguistic approximation.

118

M.S. Islam et al. / Advanced Engineering Informatics 33 (2017) 112–131

Table 2 Application of fuzzy-hybrid methods to building, roads and highways projects. Authors (year)

Method

Application area

Project

Journal/Publisher book

Reference

Thomas et al. (2006)

Fuzzy fault tree

Risk assessment

BOT road project

[45]

Wang and Elhag (2007)

Risk assessment

Bridge construction

Zeng et al. (2007) Zhang and Zou (2007)

Fuzzy logic with extension principle Fuzzy and modified AHP Fuzzy and AHP

CME/Taylor & Francis CIE/Elsevier

Risk assessment Risk assessment

IJPM/Elsevier JCEM/ASCE

[55] [59]

Wang et al. (2007)

Fuzzy integrals, MCS, AHP

Cost analysis under uncertainty

Building (shopping centre) Infrastructure (freeway construction) Building

[95]

Wang and Elhag (2008)

Risk assessment

Bridge construction

Risk assessment

Bridge construction

JCEM/ASCE

[17]

Risk assessment

Bridge construction

Risk analysis

Building construction

CME/Taylor & Francis JCEM/ASCE

[46]

Imbeah and Guikema (2009) Jin (2010) Koo et al. (2010)

Adaptive neuro-fuzzy system (ANFIS) Fuzzy set-Bayesian probability Fuzzy logic and system dynamics Bayesian probability, utility theory Fuzzy set, ANN CBR hybrid model

CAJCIE/WileyBlackwell ESA/Elsevier

Freeway projects Housing (building) projects

Fuzzy MCS

Xu et al. (2010a) Xu et al. (2010b)

Fuzzy synthetic evaluation

Risk allocation modelling

Infrastructure (highway overpass) Infrastructure (highway)

JCCE/ASCE CJCE/NRC Research Press CAJCIE/WileyBlackwell JCEM/ASCE

[89] [61]

Sadeghi et al. (2010)

Risk allocation decision making Project time and cost prediction considering risks Risk assessment and cost analysis

Yazdani et al. (2011)

Fuzzy synthetic evaluation Fuzzy COPRAS

Risk assessment modelling Risk analysis

Infrastructure Rail transportation

[34] [85]

Farajian and Cui (2011)

BBN and utility theory

Mao and Wu (2011)

Fuzzy arithmetic and fuzzy real option FAHP

financial utility assessment, facilitate decision making Risk analysis for investment decision making Risk assessment

Transportation infrastructure Real estate

AC/Elsevier EI/Taylor & Francis JCCE/ASCE SEP/Elsevier

[19]

Building rehabilitation

IJPM/Elsevier

[2]

FAHP

Risk assessment

JCEM/ASCE

[54]

Idrus et al. (2011)

Fuzzy expert system

ESA/Elsevier

[36]

Lin and Jianping (2011) Karimiazari et al. (2011) Sonmez (2011)

F-ANP Fuzzy TOPSIS ANN, Bayesian probability

Building and infrastructure Road construction Building projects

SEP/Elsevier ESA/Elsevier ESA/Elsevier

[76] [80] [60]

Yazdani et al. (2011)

Fuzzy COPRAS

Risk analysis and cost contingency estimation modelling Risk assessment Risk assessment model selection Risk relationship assessment, cost range estimation Risk analysis

Infrastructure (expressway) Building and infrastructure

Critical Infrastructure

[85]

Maravas and Pantouvakis (2012) Aboushady et al. (2013) Aboshady et al. (2013)

FST

Choi and Mahadevan (2008) Nasirzadeh et al. (2008)

Nieto-Morote and RuzVila (2011) Li and Zou (2011)

[28]

[63]

[64]

[98] [33]

[65]

Cost analysis addressing risk and uncertainty Risk analysis Risk assessment

Road construction

Ekonomska Istrazivanja IJPM/Elsevier

Building Real estate

IEEE Xplore AEIC/ASCE

[26] [38]

Risk assessment

Commercial building

PCS/Elsevier

[79]

BBN

Seismic and aging risk assessment

Bridge projects

[67]

Cheng and Hoang (2014) Khodakarami and Abdi (2014) Nguyen et al. (2015) Pawan and Lorterapong (2015) Zhao et al. (2015) Valipour et al. (2015)

Fuzzy logic, SVM BBN

Risk analysis Risk network, cost analysis

Bridge projects Hospital building

Working paper/ UBC JCCE/ASCE IJPM/Elsevier

Fuzzy AHP FST

Transportation projects High-rise commercial building Green building Freeway PPP

IJPM/Elsevier JCEM/ASCE

[7] [20]

Fuzzy Synthetic Evaluation FANP

Project complexity assessment Risk analysis and time contingency estimation Risk assessment Risk prioritisation model

[31] [93]

Ameyaw et al. (2015)

Fuzzy synthetic evaluation

Risk assessment

Abdollahzadeh and Rastgoo (2015) Salawu and Abdullah (2015)

Fuzzy logic, fault tree, event tree Fuzzy synthetic evaluation

Risk assessment

Government funded infrastructure Bridge construction

JCP/Elsevier JCivEM/Taylor & Francis JFM/Emerald Insight ASCE-ASME JRUES

Risk assessment for contractor selection

Highway projects

PSBS/Elsevier

[35]

Tamosaitience et al. (2013) Tesfamariam (2013)

Fuzzy consensus Fuzzy expert system, FTA, ETA Fuzzy TOPSIS

Cho et al. [110] introduced the fuzzy concept to address uncertainties in their method of estimating a budget range influenced by the degree of risk. Baloi and Price [108] used FST in modelling global risk factors affecting project cost overruns, finding it a feasible

[21]

[43] [66]

[32] [3]

technique for risk assessment. Dikmen et al. [39] developed a risk rating model by FST that can evaluate cost overrun risk in a project. They argue that computational complexity may be the reason for the lack of widespread use of FST in practice. Malek et al. [22]

M.S. Islam et al. / Advanced Engineering Informatics 33 (2017) 112–131

developed a FST-based model to assess the risks of concrete structures under uncertainties, where the fuzzy weighted mean represents the risk-score for finding the critical risks. Salah [24] presented a comprehensive risk management model using FST, demonstrating the application of FST in project risk assessment along with risk mitigation, monitoring and control. Fuzzy set theory provides an opportunity of using both numerical and qualitative approaches for enhancing decision analysis under uncertainty. However, an axiomatic framework is required to encode the linguistic expression into numerical scale in a meaningful way. Its criteria aggregation process, preference relations for ranking and computation methods are also identified as problematic issues. The validation technique of fuzzy based decision analysis, which is critical in establishing the model for practical implication, is also disregarded in the studies [106]. In addition, FST-based risk assessment models have a basic drawback in not providing realistic risk assessments because of sometimes omitting critical risks due to an inability to capture all possible risk scenarios when experts provide risk evaluation judgments [17]. 3.2. Extended fuzzy methods 3.2.1. Fuzzy arithmetic Fuzzy numbers can be manipulated by the usual arithmetic operations, such as addition, subtraction and multiplication, which are called fuzzy arithmetic. Two methods used in fuzzy arithmetic computation are, for example, the a-cut method and extension principle method [18]. The a-cut method can lead to the over estimation due to its interval arithmetic [111], while the extension principle method (i.e. a point-wise calculation between input fuzzy numbers and calculation of the final fuzzy numbers as the membership degree of output points) provides a more accurate estimation of uncertainty by using different t-norms [18]. Elbarkouky et al. [18] used the fuzzy-arithmetic operation for developing cost contingency estimator software by considering the probability and impact of risk on a fuzzy linguistic scale. Fuzzy arithmetic was also applied by Sadeghi et al. [101] for capturing uncertainties in measuring queue performance in construction projects. 3.2.2. Fuzzy synthetic evaluation Fuzzy synthetic evaluation is an applied form of FST that uses multiple criteria to evaluate an object relative to an objective in a fuzzy environment [30]. It has the advantage of dealing with complex evaluations having multiple levels and attributes [31]; can assess the risks from ambiguous and imprecise data elicited by subjective judgments; and can measure the level of an individual risk, group risks and overall project risk. The basic steps of fuzzy synthetic evaluation are: (a) defining a membership function of the level of an individual risk’s likelihood or occurrence as the percentage frequency of the response; (b) multiplying the membership functions by the respective weight of qualitative terms given by experts to obtain the risk likelihood or occurrence score; (c) calculating the risk index as the square root of the product of the risk likelihood of occurrence and risk magnitude; and (d) using the defined scale of the risk index to obtain the risk level [31]. However, the method is unable to cope with the randomness and discrete characteristics of project risks. It also disregards the causal relationships of the risks at different hierarchy levels. 3.2.3. Fuzzy expert system and Mamdani inference Fuzzy expert systems provide an easy way of dealing with fuzzy sets [36], situations involving both nonlinear and uncertain characteristics [112,113] and illustrating the experts’ practical way of assessing risks. The outcomes are based on expert judgment, qualitative assessment, causal relationships and impact analysis [37]. They have three basic components: a fuzzy

119

membership function, fuzzy rules and a fuzzy inference mechanism. The later typically involves five steps of ‘‘fuzzification, rule evaluation, implication, aggregation and defuzzification” [36]. Mamdani inference (see Jin [114]) is the improved, most common method and well-accepted fuzzy expert system. It is a simple ‘‘minimum operator” that reflects fuzzy ‘‘if-then” rules to obtain fuzzy conclusions from fuzzy inputs and inference processes [9,115]. 3.3. Fuzzy hybrid methods 3.3.1. Fuzzy comprehensive evaluation and cloud model While fuzzy comprehensive evaluation (FCE) can handle uncertainty and vagueness, the cloud model is appropriate for cases involving discreteness and randomness. If project risks are vague, fuzzy and uncertain as well as discrete and random by nature, and include data elicited from domain experts, then the combined FCE and cloud model provides reliable results for evaluating and prioritising risks [40]. The basic components of FCE are the risk value index, pair wise comparison between risks using AHP for evaluating their weight, fuzzy weighted average and risk evaluation based on the corresponding risk score interval. The judgment biases that can result from using the maximum degree of membership are countered by generating a cloud for risk weights and risk values by a cloud model [40]. The cloud model is based on an ‘evaluation cloud’ (containing risk weights and risk values), with components comprising the sample mean, sample variance, entropy and excess entropy of the cloud [116]. A ‘remark cloud’ is established and a decision is taken for an individual risk and group risks. The basic limitations of this model are that it does not consider the impact of risks on each other within the group or beyond the group, nor any complex relationships of the risks. Thus, if some of the risks are structured (having causal relationships) in nature, the model will not be appropriate for risk assessment. 3.3.2. Fuzzy fault tree and fuzzy event tree analysis Fuzzy fault tree analysis (FTA) and fuzzy event tree analysis (ETA) are probability-based risk assessment tools that have been used for solving MCDM problems. FTA has the advantage of providing a good sketch of the root-causes of risks, making it easy to visualise and understand an event with imprecise information, although very complex project risk relationships are impossible to capture in this way [44]. FTA is a graphical model of some parallel events under some compound events and a series of basic events leading to the occurrence of an unexpected event, called a top event [3]. For FTA, the system is first defined and a fault tree structure is constructed using an ‘‘AND” or ‘‘OR” gate, as a mediator to find the upper event from the combined effects of lower events. A minimal cut set is then measured qualitatively using Boolean algebraic analysis of a fault (basic event). Finally, a quantitative analysis is carried out to calculate the probability of occurrence of the top event. The basic drawback of this method is in obtaining an accurate estimation of the occurrence probability, although this may be overcome by adding fuzzy logic to improve the model [3]. This fuzzy FTA provides the fuzzy occurrence probability of the top event and offers fuzzy linguistic options for experts to assess the probability of basic events [44]. The ETA is also a graphical model, where an initiating event is identified and defined, pivotal events identified and an event tree constructed. The probability of an initiating event is determined, and the binary (0, 1) probabilities of success or failure of the pivotal events are also determined. The overall probability of occurrence of each scenario is then obtained from the initiating event and pivotal events. Here, the probability of occurrence of the final event, success and failure of an event or a scenario of occurrence is a crisp value. The accuracy of the ETA model’s outcomes depends on the

120

M.S. Islam et al. / Advanced Engineering Informatics 33 (2017) 112–131

accuracy of the information provided. However, the use of ETA is limited, as project risks are uncertain and vague in nature. Thus, introducing fuzzy logic improves the ETA model by considering the probability of occurrence of the final event and the success/failure probability of pivotal events as a fuzzy, instead of crisp, number. It also performs fuzzy arithmetic operations in ETA [3]. 3.3.3. Fuzzy and Monte Carlo Simulation (FMCS) Monte Carlo simulation (MCS) is a decision making tool that depends on the probability of occurrence of an event from historical records. The impact of construction risks depends on numerous factors (e.g., human activities, workplace conditions, and nature and availability of resources) with most being represented in linguistics terms by subjective judgments. Probabilistic analysis (i.e. MCS) of the linguistic terms for evaluating risks has been critiqued because the subjective judgment of an individual may not be suitable for precise scientific inference [117]. Fuzzy logic is best at handling subjective and linguistically expressed data and also provides a solution [62] by combining both quantitative and qualitative data in conjunction with a simulation technique [118]. Such a fuzzy hybrid approach – for example, fuzzy calculus with the acuts method and Monte Carlo random sampling for probability distribution functions - has been proposed by Guyonnet et al. [118]. The main shortcomings of these approaches are that the Infimum (Inf) and Supremum (Sup) values cannot always represent the a-cuts of a fuzzy set; there is no mention of the reason for using Infimum and Supremum values of the output a-cut from the 5% probability of the lower and higher values of the a-cuts histograms; and that the outcome of this model will be different from the traditional MCS method in considering only random inputs instead of fuzzy inputs as an extreme condition [98]. The absence of fuzzy inputs will not provide enough probabilistic information for decision making under uncertainty. Sadeghi et al.’s [98] alternative approach consists of a novel methodology for the simulation of input data with fuzzy and probabilistic uncertainty by introducing a fuzzy-MCS approach for risk assessment and cost-range estimation. However, the approach lacks a suitable method for fuzzy random generation to produce suitable sample sets. It uses a probability-possibility method to transform some of the probability distributions into fuzzy sets, and applies fuzzy arithmetic to calculate the output as a fuzzy set based on the Inf and Sup of the input a-cut intervals. Another simulation-based fuzzy hybrid approach called fuzzy randomness-MCS, has been presented by Attarzadeh et al. [100] to analyse the linguistic expression of risk evaluation found by subjective judgment capturing the uncertainties and vagueness in risk assessment. Similar to Sadeghi et al. [98] they also developed a fuzzy cumulative distribution function (CDF), which provides a range estimation of decision variables to facilitate decision-making under uncertainties for complex projects. However, the fuzzy-simulation approach is not easy or straightforward for complex projects. In addition to the absence of sufficient data, this simulation method is also unreliable for decision analysis and is incapable of presenting the causal dependencies among risk factors [119]. 3.3.4. Fuzzy Artificial Neural Network (F-ANN) The ANN is an artificial intelligence-based nonparametric model that has been used for risk analysis. ANN has the capacity to be trained from past data and be applied to generating a future outcome [75,76]. The ANN process involves a bunch of simulated neurons (processing elements) unified in such a way that the neurons are capable of being trained [120,121]. ANN often produces more accurate results than other conventional methods (e.g., regression analysis) and is ideal in situations where there is lack of information regarding risks, and a complex, nonlinear or unknown relationship exists between project risks [87,88,122]. However,

the ANN model provides a single value rather than a range to define project risks [60]. It has a hidden layer, which is unable to clarify the model’s structure [61]. In ANN, the processing elements need to be trained properly by historical data from similar projects, which mostly depends on the performance of the neurons and the availability of sufficient data [60]. Individual construction projects are unique in nature and only a limited risk data is available for the evaluation of project risks [98]. As mentioned earlier, the lack of sufficient data usually means that analysts have to rely on expert judgments and the linguistic expression is often an easy and more natural way to evaluate risks from such judgments [44,63]. However, uncertainty exists in expert judgment due to ambiguity, vagueness, ignorance and imprecision in understanding and evaluating the project risks [47]. This uncertainty cannot be addressed by the traditional ANN model [123] and fuzzy logic is often used to address this application gap with ANN [63]. In their review work, Chan et al. [11] discovered that the combination of fuzzy and ANN would be a potential tool for uncertainty modelling in construction risk management. 3.3.5. Fuzzy failure mode and effect analysis Failure Mode and Effect Analysis (FMEA) is a tool for identifying potential modes of failure in a system, evaluating the main causes, determining the impact of failures and formulating preventive measures [50]. In this system, a Risk Priority Number (RPN) for each failure mode or risk event is computed as the product of the probability of risk occurrence (O), severity (S) and detection (D). The RPN represents the level of a particular risk, i.e. a higher value of RPN means higher level of risk. This rating has been used in different studies to find critical risk factors [51]. In FMEA, ‘‘detection” is an important term exemplified as the capability of identifying the risk of not having enough working hours to take corrective action. The introduction of ‘‘detection” in prioritising and assessing risk provides a new dimension to reach a higher level of accuracy. However, the FMEA technique assumes that S, O and D are equally important [124], which is not always realistic [56]. It is also difficult to determine the precise probability of a failure event in FMEA if the data is linguistic [125]. The application of fuzzy logic in FMEA can address these drawbacks, in that the experts have the opportunity to assess O, S and D in linguistic form. The fuzzy FMEA also provides an easy but efficient mechanism for modelling project risk assessment [51]. 3.3.6. Fuzzy TOPSIS In a fuzzy environment, the Technique for Order of Preference by Similarity to Ideal Solution (TOPSIS) is a new method that is very suitable for project selection, bid and risk evaluation for risk assessment by multiple criteria decision analysis (MCDA) [8]. TOPSIS is a matrix method that provides a suitable and easy way of computing the weights of alternatives based on similar preferences, but it is incapable of handling any uncertainty and vagueness in expert responses [126]. Alternatively, the fuzzy method can handle uncertainty and vagueness, but it provides only single value of risk as the outcome, which is not always appropriate for reaching decisions because of some information gap. In such a case, adding TOPSIS with fuzzy can solve this problem [8,82]. The combined fuzzy TOPSIS method can also handle both qualitative and quantitative data, and provide the outcome in quantitative form for project risk assessment [82]. The method defines the fuzzy weights of risk evaluation criteria (i.e. time, cost, quality and safety) based on expert judgment. It forms alternatives and a criteria matrix for each expert and applies a max-min rule for fuzzy inference decisions taken from the opinion of multiple experts. The matrix is then transformed into a fuzzy weighted normalised decision matrix, and positive/negative Euclidian distances and closeness coefficients are measured for all alternatives with respect

M.S. Islam et al. / Advanced Engineering Informatics 33 (2017) 112–131

to each risk. Finally, the efficiency rating for each alternative is determined for ranking the alternatives [8]. 3.3.7. Fuzzy VIKOR The VIKOR method is considered as ‘‘one applicable method” in MCDA [127] used for group decision-making to select the best project option from alternatives with the same evaluation criteria. It is an extended form of linear programming metric, in which ranking is performed using the closeness of a particular measure to an ideal alternative in comparing different alternatives. The VIKOR method allows the ranking of different alternatives using the values of different evaluation criteria weights, and analysing the impacts of criteria weights on a proposed alternative. The method has some advantages for use in decision analysis. For example, it is easy and stable to use with cardinal data and considers the lowest performance rating for a specific attribute [127,128]. It is a suitable tool in multicriteria decision making, specifically for the decision makers who are unable to express their judgments for risk evaluation at a preliminary stage [127]. The method provides maximum group preference and minimum individual regret (the opposite to preference) to evaluate project criteria. However, the VIKOR method is not a suitable tool for evaluating project alternatives if the evaluation criteria are uncertain, vague and imprecise. In such cases, a modified VIKOR method with fuzzy logic or FST has been used [129,130]. 3.3.8. Fuzzy Analytical Hierarchy Process (FAHP) The AHP is a predominant MCDM technique in risk and uncertainty analysis and one of the best methods for measuring project complexity [48,131]. However, it cannot cope with inconsistencies in pairwise comparisons [2]. The combination of fuzzy logic and AHP is recognised as the most influential risk management method [29]. The fuzzy-AHP (FAHP) model can effectively measure subjective data under multiple divergent risks in a project [29]. Zeng et al. [55], for example, combined AHP with fuzzy logic to rank risk preferences from expert judgments. The model determines risk magnitude by aggregating the risk factor index, probability and intensity of risk into a fuzzy decision system. However, FAHP method has some limitations. For example, AHP does not indicate the causal relationships between the risk factors at the same level and does not consider the impact of risks in different phases of the project life-cycle, and is therefore impractical to use with a large number of risk evaluation criteria [77] where an enormous and tedious number of pairwise comparisons are needed. Hence, the FAHP method needs to be improved by other optimisation tools. 3.3.9. Fuzzy Analytical Network Process (FANP) As MCDA methods, the AHP and Analytical Network Process (ANP) are frequently used for risk analysis. The AHP considers risks as independent elements of a hierarchy structure [92], while the risks in complex projects are highly interdependent. In contrast, the ANP captures the interdependencies and impact of various risks for risk ranking [91]. Similar to the AHP, ANP considers pairwise comparisons of the risks, but unlike AHP, it captures all the possible causal relationships and networks between the clusters (group-risks) and among the elements (sub-risks) within a cluster [132]. However, ANP assigns a crisp value in the pairwise comparison of risks, which is a limitation to capturing vagueness and uncertainty in risk analysis. Thus, introducing fuzzy concepts in ANP provides an advanced step in overcoming this limitation [93]. The fuzzy-ANP (FANP) first identifies all potential risks and their interdependencies and builds a network model showing their causal relationships. It then makes pairwise comparisons between the risks using a suitable fuzzy linguistic scale or fuzzy number, tests for consistency between the data sets, aggregates the judgment matrices, calculates priority weights, computes and limits

121

the super-matrix, and ranks the risks based on the calculated scores [92]. As with the AHP, however, assessing project risks based on many criteria requires an enormous number of pairwise comparisons. 3.3.10. Fuzzy Bayesian Belief Networks (F-BBNs) The BBN can handle complex and uncertain relationships in risk networks [66,70,71,73]. It is graphically defined by a directed acyclic graph (DAG) where nodes represent risks and arrows represent the causative relationships between the risks. The arrows also denote the uncertainties inside the risk network, which are mutually inclusive under the concept of conditional probability [133,134]. Using the BBN, large and complex risk networks can be easily constructed by the aggregation of sub-networks into hierarchy levels [68,135]. Two types of probabilistic data are required for any Bayesian network, the prior probability of independent risks and the conditional (effect as the influence of cause) probability of dependent risks given that of the independent risks. Using Bayesian probability theory, it is then easy to obtain the probability of a dependent risk [136]. By placing risks in hierarchy levels, this network reduces the need for collecting a huge amount of data, as it helps in computing the probabilities of upper level of dependent risks from the probabilities of lower level risks (prior probability) and their probabilistic dependencies (conditional probability) [66,73]. Additionally, the method is a powerful tool for working with an inadequate and small number of datasets, data found from a mixture of different areas of knowledge, non-parametric and distribution-free data, and data for a highly diverse set of variables [66,70,71,137]. Moreover, a BBN can easily update the probabilities in the network when new data for the variables becomes available [68,71,138]. It can also deal with the prediction, deviation detection and optimisation of decision variables based on very subjective judgments [69]. Bayesian networks assess the reliability, vulnerability and effectiveness of a system using the probabilities of the variables and the stochastic dependencies among the variables [137]. It computes the risk detectability and probability of false detection for assessing system reliability [139]. In comparison with other risk assessment methods such as ANNs, MCS, CBR and system dynamics, the BBN has a great advantage in its simplicity for use by practitioners and its accuracy with respect to the amount of data available [60,71,140]. In BBN analysis, very precise data is required for the prior and conditional probabilities, which is difficult to obtain from large and complex projects because of the amount of uncertainty involved [74]. There is also a lack of sufficient data for the risk assessment of complex construction projects, which leads to having to rely on expert opinion for data elicitation. In addition, while expert judgment is required to develop BBNs, there is limited research on how to elicit knowledge from the experts and ensure the reliability of the model [136]. Fuzzy logic helps domain experts to express the frequency and consequences of risk linguistically, which can be transformed into a range or PERT-like three-point probability (low, medium and high). However, fuzzy logic alone cannot express the causal relationships between the risks and is unable to conduct inverse inference [17,72]. Thus, a combination of fuzzy logic and BBN theory (i.e. FBBN) has a significant role to play in expediting project risk analysis in an uncertain environment [141,142]. 4. Applications of fuzzy and hybrid methods to construction risk management Fuzzy and hybrid methods have long been used in different areas of construction risk management studies [11]. The applica-

122

M.S. Islam et al. / Advanced Engineering Informatics 33 (2017) 112–131

tion areas of fuzzy and hybrid methods are classified into risk assessment, time and cost analysis under uncertainty, and risk assessment for decision making (project planning, selection and bid evaluation) [11]. In order to validate or demonstrate the applicability of the developed models and methods, extant studies have used a wide range of construction projects, such as buildings, roads, highways, expressways, freeways and bridges, pipelines, subway tunnelling, power plant, transmission lines and real estate. The following sections discuss the reported applications of fuzzy and hybrid methods in the literature for the risk management of different civil engineering projects (Tables 2–4).

et al. [38] presented a comprehensive risk management strategy based on a fuzzy expert system (FES), FTA and ETA in detecting and assessing critical risks of real estate projects. In their approach, risks are ranked by FES, while FTA and ETA respectively compute the probabilities of critical risk events and provide mitigation measures. Table 2 summarises recently published articles concerning the risk and uncertainty analysis of building and infrastructure projects using fuzzy and hybrid methods.

4.1. Building and real estate projects

Build-Operate-Transfer (BOT) is one of the most popular arrangements for procuring major infrastructure projects such as roads and highways. BOT projects however are considered as complex, high-risk projects due to the large capital investment involved, long-term contracts and payback period, and other complexities. Thomas et al. [144] proposed a fuzzy fault tree model for the risk assessment of road construction projects, with fuzzy sets being used to define risk probabilities from expert judgments. Joint venture contracts (JVC) between the host country and overseas parties provide another important but complex means of financing and developing infrastructure projects. Zhang and Zou [59] presented a FAHP model for the risk assessment of JVC freeway construction projects; while Wang and Elhag [28] developed a fuzzy group decision making (FGDM) model for JVC bridge construction projects that involves solving the fuzzy weighted average by a linear programming model instead of fuzzy arithmetic to provide a more exact risk assessment. Similarly, Wang and Elhag [63] used an adaptive neuro-fuzzy system (ANFIS), Choi and Mahadevan [17] a fuzzy set and Bayesian probability method, and Nasirzadeh et al. [46] fuzzy logic and system dynamics in developing risk assessment models for JVC bridge construction projects. Infrastructure projects procured by PPP arrangements face numerous risks owing to the long term contracts and huge investment involved [54]. For PPP infrastructure (highway) project risk assessment, Xu et al. [34] developed a fuzzy synthetic model through an empirical study to quantify critical risks and overall project risk based on objective evidence instead of subjective judgment; Li and Zou [54] proposed a FAHP model for risk assessment of an expressway project; while Yazdani et al. [85] used fuzzy complex proportional assessment (FCOPRAS) for risk analysis of PPP infrastructure projects to determine the criteria weights and importance of alternatives. Valipour et al. [93], on the other hand, developed a risk prioritisation model called ‘‘fuzzy analytical network process (FANP)” for freeway PPP projects to facilitate risk management during the preconstruction stage in which the project’s risks are identified mostly by subjective judgment. The nature of uncertainties and risks in PPP projects, particularly PPP-BOT type projects, is such that a definitive project cost forecast is often unrealistic. Rather, cost range estimation with risks is more natural to facilitate decision making in such projects. Attarzadeh et al. [100] found that the traditional MCS for cost estimation is not suited to capturing project risks and uncertainties, and consequently presented a fuzzy randomness-MCS model for a toll road and bridge project. In this model, fuzzy-CDF of cost estimation provides a range of estimation, thereby facilitating the negotiation price of the project under uncertainties. Another popular approach to risk management involves setting the contingency amount as a percentage of the estimated cost that is largely based on subjective judgment, and which is difficult to justify. On the other hand, more sophisticated methods use complex simulation and modelling techniques to apportion the contingency amount, but are impractical for infrastructure projects [36]. An alternative approach, adopted by Idrus et al. [36] for infrastructure projects, involves the use of a fuzzy expert system (FES) model for

As noted by Wang et al. [95], some risk factors have a direct influence on project cost and are treated as cost uncertainty in the estimation process. They developed a cost estimation model integrating MCS with fuzzy integral and AHP to capture risks and uncertainties. In this model, fuzzy integral, which is a fuzzy technique for aggregating the values of risk evaluation criteria [143], was used for assessing sub-risk factors. MCS was used to generate random costs of the direct and indirect cost components, and AHP for measuring the comparative weights of the risk factors. Zeng et al. [55] adopted a fuzzy reasoning technique and modified AHP as a risk assessment method and justified its use for a building project; while Nieto-Morote and Ruz-Vila [2] presented a FAHPbased risk assessment method, with an illustrative example of risk assessment of a building rehabilitation project. The fuzzy concept in these fuzzy hybrid methods is used to handle the uncertainties and biases in a subjective data set, and AHP for developing a hierarchical structure of a large number of risks and pairwise comparisons between the risks, with a view to handling the inconsistencies of expert judgments [55,95]. Aboushady et al. [26] have developed a ‘‘Fuzzy Consensus Qualitative Risk Analysis Framework” for evaluating risks in building projects by using FST and Euclidean Distance Measurement systems. Three different fuzzy algorithms were used for aggregating expert opinions, such as the ‘‘Fuzzy Similarity Aggregation Method”, ‘‘Fuzzy Distance Measurement Method” and ‘‘Fuzzy Optimal Aggregation Method”. Of these three methods, the ‘‘Fuzzy Optimal Aggregation Method” was found to be the most accurate in qualitative risk management. Zhao et al. [31] presented a risk assessment model for green building projects using fuzzy synthetic evaluation – an applied form of FST – to allow mathematical operations in the fuzzy domain and apply quantitative risk analysis to qualitative data. Pawan and Lorterapong [20] measured time contingency by applying expert judgments, using FST to capture the imprecision and vagueness in risk assessment and considered the impact of multiple risks in estimating activity duration. The method was demonstrated for a high-rise commercial building and found to be realistic for project scheduling. The method provides an agreement index between the fuzzy estimated duration and contract duration of a project to facilitate decision-making. Tamosaitience et al. [79] developed a risk evaluation and ranking model of expert opinions for selecting a construction project from different alternatives using a fuzzy TOPSIS method and validated the model for commercial building projects. In their model, fuzzy logic was used to quantify linguistically expressed risk criteria on a numerical interval scale, which improved the TOPSIS weight determination process and provided the best result for decision making [79]. Fuzzy and hybrid methods have also been applied to risk analysis of real estate projects (Table 2). For instance, Mao and Wu [19] used fuzzy arithmetic and fuzzy real option parameters in assessing the potential risks of real estate projects. The aim is to support investment decisions by evaluating project value and addressing the risks involved in project cost and expected revenues. Aboshady

4.2. Roads and highways projects

M.S. Islam et al. / Advanced Engineering Informatics 33 (2017) 112–131

cost contingency estimation. Like cost range estimation and cost contingency modelling, risks and uncertainties also need to be considered in project cash flow analysis (i.e. analysing project accrual costs and corresponding payments with respect to time) for the project execution phase. Thus, researchers have developed project cash flow analysis models under uncertainties based on fuzzy methods. For example, Maravas and Pantouvakis [21] applied FST to analyse the cash-flow scenarios of a road construction project under potential risks by considering the individual task durations and costs of a project in activity networks. However, a major drawback of their study is the assumption of deterministic indirect costs with respect to time or ambiguous situations. Jin [89] developed a novel risk allocation model based on artificial neural networks (ANNs) and fuzzy inference systems (FISs) for PPP road construction projects. They argue that probabilistic models may provide unrealistic results, as they do not consider non-linear relationships of the risk factors. In contrast, ANFIS (i.e. combination of ANN and FIS) is capable of handling the uncertainty, vagueness, complexity and nonlinearity of project risks [89]. Cheng and Hoang [43] presented a model combining the fuzzy logic, least-squares method and support-vector machine for prioritising risks to support decision makers in bridge maintenance planning. They applied the fuzzy logic to enhance an approximate reasoning capability and to deal with subjective information. In contrast, Abdollahzadeh and Rastgoo [3] assessed the risks of bridge construction projects using fault tree and event tree analysis. As infrastructure projects often fail to be completed on time and within the budgeted cost because of numerous uncertain events [145], better management of project risks by the contractor can help reduce the impact of these uncertainties [146]. The selection of an appropriate contractor based on risk management capability and other criteria is a typical MCDA problem. Salawu and Abdullah [35] used FST with a four level scale to assess contractor risk management maturity for highway infrastructure projects. 4.3. Subway/tunnel projects Subways involve complex construction in urban areas to solve increasing transportation problems. There has been an increased level of enthusiasm and interest from researchers in such projects due to the multitude of risks and uncertainties involved [53]. In the early design stage, AHP can be used to breakdown subway projects into different levels of subsystems to understand their project risks and uncertainties. Zou and Li [53] proposed FAHP to facilitate risk identification and assessment; while Fouladgar et al. [81] employed a fuzzy TOPSIS framework that includes consequences, detectability, vulnerability and reaction against an event. The term ‘‘vulnerability” is overlooked in conventional risk evaluation models, although any kind of weakness in the project life cycle that can be turned into a potential active project risk [81]. Mohammadi and Tavakolan [50], on the other hand, proposed a risk assessment approach based on fuzzy FMEA, validating the approach on a subway construction project. This views the impact of each risk as the aggregation of time, cost, quality and safety, with inferential fuzzy ‘‘if-then” rules applied depending on likelihood, impact and detection (the basic components of FMEA). Most recently Yan et al. [94] also commented that tunnel projects are very complex, with risks that are uncertain, random, interrelated and fuzzy by nature. In response, they developed a fuzzy analytical network process (FANP), in which a fuzzy comprehensive assessment is used to address the uncertainties and fuzziness in experts’ judgment and ANP is applied to model the causal relationships between the risks at the same or different levels. Cárdenas et al. [136] used Bayesian Belief Network (BBN) to facilitate risk management in tunnel construction, identifying judgment bias as one of the key issues overlooked by previous

123

studies, and important to address in BBN models to ensure reliability of the model’s outcome. They further recommended modelling uncertainty by applying the observed information from project advancement. In a further study, Cárdenas et al. [147] developed probabilistic causal models based on Bayesian networks and investigated their use in identifying critical tunnel risks. The most challenging issue of risk assessment in tunnel projects, according to the authors, is obtaining information relating to tunnelling risks, which are both project and context dependent. Špacková et al. [148] presented Dynamic Bayesian Networks (DBN) to estimate tunnel construction time, demonstrating the process of continuously updating construction time prediction based on observed performance data. A limitation is that they use a single value human factor (although it can be randomly varied over the construction period). Straub [149] proposed a reliability estimation model using Bayesian updating techniques that was later applied to geotechnical construction [150]. Camós et al. [151] developed a reliability based method for predicting the probability of the damaging effect of tunnel construction on other nearby structures. The reliability approach was based on the Bayesian probability technique, which allows the model to be updated in response to observed data during construction. Compared with Monte Carlo Simulation, this method is able to cope with the knowledge of dynamic changes of construction parameters and underlying uncertainties. A similar probabilistic method (i.e. Bayesian updating model) for the prediction of tunnel induced settlement of buildings and other nearby structures was also presented by Camós et al. [152]. They used reliability-based criteria for modelling the probabilities of the settlement of structures caused by tunnel work and which updates the predicted settlements when new measurements become available. Since the Bayesian probability technique requires the exact probability value of an event or risk, which is very critical information but can be hard to obtain with precision, Zhang et al. [74] developed a fuzzy Bayesian networks model to determine the safety risk analysis of pipeline damage induced by tunnel construction. They quantified the probabilities of safety risks based on both previously observed variables (i.e. quantitative data) and subjective judgment of some other variables. They introduced a confidence-based approach to assess the experts’ judgment reliability in determining the fuzzy probability of occurrence of events. 4.4. Pipelines projects Fares and Zayed [9] used a hierarchical fuzzy expert system (HFES) to evaluate the risks of water main failure. They claimed that the fuzzy logic-based method was appropriate for summarising fuzzy linguistic expert knowledge, addressing imprecisely defined problems and making reasonable risk management decisions or educated guesses in uncertain situations. Abdelgawad and Fayek [51] developed a fuzzy expert system in a traditional FMEA system for pipeline projects. They applied the system to establish the relationship between three basic elements of FMEA i.e., the probability of occurrence (O), severity risk (S) and detection level to control the risk (D). A FAHP was then conducted to determine the aggregated impact of risk from the three basic elements. In an effort to overcome the drawbacks of traditional FMEA for dealing with uncertain and inadequate data with complex pipe jacking projects, Cheng and Lu [10] proposed a fuzzy FMEA model for risk assessment. In a similar vein, Abdelgawad and Fayek [44] adopted a risk assessment framework for pipeline projects using fuzzy FTA. FTA is a technique for locating the root causes of a system failure or risk by the probability of occurrence of a risk and is used in their study by eliciting fuzzy probability from experts to obtain a more reliable result rather than a crisp probability from imprecise data. They proposed a comprehensive risk management framework incorporating the combination of fuzzy logic, FMEA,

124

M.S. Islam et al. / Advanced Engineering Informatics 33 (2017) 112–131

Table 3 Application of fuzzy-hybrid methods to subway/tunnel and pipelines projects. Author/s (year)

Method

Application Area

Project

Journal/Publisher

Reference

Fares and Zayed (2010) Zou and Li (2010)

Fuzzy expert system FAHP

Risk assessment Risk assessment

Water mains Subway construction

[9] [53]

Abdelgawad and Fayek (2010) Abdelgawad and Fayek (2011) Abdelgawad and Fayek (2012) Fouladgar et al. (2012) Mohammadi and Tavakolan (2013) Jamshidi et al. (2013) Kuo and Lu (2013)

Fuzzy FMEA, FAHP

Risk criticality assessment

Pipeline construction

JPSEP/ASCE CME/Taylor & Francis JCEM/ASCE

Fuzzy FTA

Assessment of risk events

Pipeline construction

JCEM/ASCE

[44]

Fuzzy logic, FMEA, FTA, ETA Fuzzy TOPSIS Fuzzy FMEA

Risk analysis

Pipeline construction

JCEM/ASCE

[52]

Risk evaluation Risk assessment

Tunnel construction Subway construction

ACME/Elsevier IEEE Xplore

[81] [50]

Fuzzy Mamdani Fuzzy MCDM

Risk assessment Risk assessment

JLPPI/Elsevier IJPM/Elsevier

[14] [86]

Cárdenas et al. (2013)

BBN

Risk management

Pipeline project Underground construction Tunnel construction

[136]

Špacˇková et al. (2013)

Construction performance assessment

Tunnel construction

Camós et al. (2014) Cárdenas et al. (2014b)

Dynamic Bayesian networks Bayesian updating model Bayesian networks

RA/Wiley Online Library TUST/Elsevier

Settlement risk assessment Decision support for risk management

Tunnel construction Tunnel construction

[152] [147]

Cárdenas et al. (2014a)

BBN

Risk analysis

Tunnel construction

Mu et al. (2014)

Fuzzy synthetic evaluation

Subway construction

Cheng and Lu (2015)

Fuzzy FMEA

Assessing contractor risk management capability Risk assessment

TUST/Elsevier RA/Wiley Online Library RA/Wiley Online Library IJPM/Elsevier AC/Elsevier

[10]

Yan et al. (2015) Yu et al. (2015)

Risk analysis Risk evaluation

EJGE ICTE/ASCE

[94] [41]

Kabir et al. (2016)

FANP Fuzzy comprehensive approach Fuzzy-BBN

SIE/Taylor & Francis

[75]

Zhang et al. (2016)

Fuzzy-BBN

Safety risk analysis

RA/Wiley Online Library

[74]

Safety risk analysis

fault trees and event trees and applied it to pipeline projects [52]. While the outcome of the model is very similar to MCS, the use of fuzzy logic provides a new dimension where experts express their knowledge and experience linguistically to quantify risk as expected monetary value. For a similar reason, Jamshidi et al. [14] developed a fuzzy inference system for the risk assessment of pipeline projects using fuzzy logic and the Mamdani algorithm on the basis of expert judgment, with the aim of providing more accurate and precise risk assessment. As fuzzy logic does not allow the dependencies among the risk events to be modelled, Kabir et al. [75] combined the fuzzy logic and Bayesian belief network (BBN) to develop a safety risk assessment model for oil and gas pipeline projects. The fuzzy-BBN model captures the dependencies of the risk variables, uncertainties and updates the model based on probabilistic data derived from observed variables. Table 3 shows some recently published risk/uncertainty analysis articles based on the fuzzy and hybrid methods for different subway/tunnel and pipelines construction projects. 4.5. Power generation and transmission projects Table 4 lists some recently published risk and uncertainty analysis work for power generation and transmission projects. For example, Dikmen et al. [39] presented a fuzzy logic-based model for the risk assessment of power plant projects using influence diagrams to construct a risk visualisation model and applying basic fuzzy logic for rating risks. Project financing by BOT is becoming popular for power plant projects in many Asian countries. The project complexities and long-term contracts between different parties produce many risks and uncertainties in such projects. For the improved risk management of these projects, Ebrahimnejad et al. [84] presented a novel multi-criteria risk identification and

Subway pipe jacking projects Tunnel construction Subway station construction Gas-pipeline construction Tunnel construction

[51]

[148]

[71] [30]

assessment model using fuzzy TOPSIS and a fuzzy linear programming technique for multidimensional analysis of preference (FLIMAP). Their study shows that FLIMAP performs better than fuzzy TOPSIS when there are numerous alternatives. Kucukali [15], upon finding a lack of probability data, particularly for hydropower plant projects, developed a fuzzy weight rating tool for risk assessment using fuzzy logic instead of numerical probabilities for processing the linguistic judgment of experts. Ji et al. [25] also presented their fuzzy entropy-weight MCDA method for hydropower plant projects, which is an integrated method of FST, entropy-weight and TOPSIS. Here, the fuzzy decision matrix is normalised by the entropy-weight method and the fuzzy relative closeness of alternatives calculated by fuzzy TOPSIS. However, although introducing entropy-weight increases the method’s objectivity by turning the uncertainties into fuzzy sets and disregarding subjective weights, it is unable handle the uncertainties in random data. Zhao and Li [40] overcame this by a combined fuzzy comprehensive evaluation (FCE) and cloud model for UHV power plant projects, where the cloud model takes care of data randomness and the uncertainty in risk data by the FCE method. 4.6. Other construction projects The combined fuzzy and AHP for risk management in construction projects features regularly in many articles. For example, Chen and Wang [57] developed a risk assessment model for international construction projects using a combination of fuzzy logic and AHP, where fuzzy logic assesses the risk scores and AHP provide the risk weights. Subramanyan et al. [58] applied FAHP for the risk assessment of construction projects, and Abdul-Rahman et al. [29] used a fuzzy synthetic approach with AHP in developing a quantitative risk assessment model for qualitative and imprecise

125

M.S. Islam et al. / Advanced Engineering Informatics 33 (2017) 112–131 Table 4 Application of fuzzy-hybrid methods to power plant and other construction projects. Authors (year)

Method

Application area

Project

Journal/Publisher Book

Reference

Shaheen et al. (2007) Dikmen et al. (2007) Ebrahimnejad et al.(2008) Ismail et al. (2008)

FST

Cost analysis under uncertainty

JCEM/ASCE

[96]

Fuzzy logic

Risk assessment

Sanitary tank construction Power plant construction

IJPM/Elsevier

[39]

Fuzzy TOPSIS

Risk assessment

Gas reservoir

IEEE proceedings

[83]

Fuzzy logic

Risk identification and assessment

ANP

Risk assessment

JAS/Asian Networks for Scientific Information CJCE/NRC Research Press

[16]

Bu-Qammaz et al. (2009) Chen and Wang (2009) Ren et al. (2009) Ebrahimnejad et al. (2010) Kucukali (2011) Hsiao et al. (2012) Zegordi et al. (2012) Shafiee (2015) Ji et al. (2015) Zhao and Li (2015)

FAHP

Risk assessment

IEEE proceeding

[57]

Fuzzy BBN Fuzzy TOPSIS

Safety risk analysis Risk identification and assessment

Construction project (not specified) International construction projects International construction projects Offshore operations Power plant construction

JOMAE/ASME ESA/Elsevier

[72] [84]

Fuzzy logic Fuzzy ANN, fast messy GA Fuzzy ANP, Fuzzy TOPSIS

Risk assessment Cost analysis under uncertainty Risk assessment

Hydropower plant Semi-conductor Power plant construction

EP/Elsevier CAJCIE/Wiley-Blackwell IJET Part B: Applications

[15] [90] [82]

Fuzzy ANP Fuzzy-entropy, TOPSIS Fuzzy comprehensive evaluation, cloud model Fuzzy ANP, Fuzzy VIKOR

Risk mitigation strategy selection Risk assessment Risk evaluation

Wind power plant Hydropower plant UHV power transmission

ESA/Elsevier ESA/Elsevier Sustainability

[92] [25] [40]

Risk analysis for project selection

AMM/Elsevier

[77]

FAHP

Risk assessment

JCEM/ASCE

[58]

FST, evidence reasoning approach Fuzzy synthetic analysis, AHP

Risk management framework

Construction project (not specified) Construction project (not specified) Sea ports and terminals

ESA/Elsevier

[27]

Risk evaluation

Civil engineering

JCivEM/Taylor & Francis

[29]

BBN

Risk analysis

Offshore drilling

SS/Elsevier

[68]

FST Fuzzy Randomness, Fuzzy MCS

Risk management maturity model Time, cost, and quality optimisation Risk assessment, project selection

Construction firms Dam construction

JCEM/ASCE JCCE/ASCE

[23] [99]

Construction project (not specified) Construction project (not specified) Construction project (not specified) Underground reservoir for oil and gas

ASC/Elsevier

[8]

IEEE Xplore

[18]

KSCE JCE/Springer

[49]

JRUES Part A/ASCE-ASME

[42]

Ebrahimnejad et al. (2012) Subramanyan et al. (2012) Mokhtari et al. (2012) Abdul-Rahman et al. (2013) Khakzad et al. (2013) Zhao et al. (2013) Heravi and Faeghi (2014) Taylan et al. (2014)

FAHP, fuzzy TOPSIS

Elbarkouky et al. (2015) Jung et al. (2015)

Fuzzy arithmetic

Wang et al. (2016)

FCE

Fuzzy integral, AHP

Contingency estimation software addressing risks Cost contingency modelling by risk analysis Risk assessment

risk data. Jung et al. [49] also developed a computer-based, fuzzy integral and AHP-based cost contingency estimation process to identify the potential risks and their cost impact on public construction projects. However, the FAHP method does not consider dependencies and feedback between the risk factors and, in order to overcome this problem, Ebrahimnejad et al. [77] proposed a novel approach based on FANP and fuzzy VIKOR methods respectively to assess potential risks and compare different alternatives in selecting projects. Similarly, Taylan et al. [8] used fuzzy TOPSIS instead of the fuzzy VIKOR method, with FAHP used to produce the fuzzy linguistic variable weights, which are then applied into fuzzy TOPSIS to evaluate project risks in an incomplete and uncertain environment. The purpose is to identify a low risk project from multiple alternatives. Heravi and Faeghi [99] used MCS and a fuzzy additive weighting system for stochastic optimisation of fundamental project objectives, i.e. time, cost and quality. They use MCS simulation for the probabilistic distribution of time and cost based on observed values and fuzzy aggregation rules for quality optimisation based on expert judgements. Wang et al. [42] applied a FCE model to assess the risks of the underground construction of a crude oil reservoir. They argue that there is limited research in risk assessment of large underground storage facilities, considering soil stability and the safety of containment. They used expert opinion to assess the likelihood and consequence of risks to calculate

[91]

the risk score. The method, however, does not capture the causal relationships and dependencies between the risks. Mokhtari et al. [27] developed a FST based framework to evaluate the risk factors in response to uncertainty and used evidential reasoning (ER) to synthesis the outcome for seaport and terminal projects. Other applications of fuzzy and hybrid methods are: a FSTbased cost estimation model for an unusual large sanitary tank construction project, with insufficient probabilistic data for cost estimation by MCS, leading to the need to rely on expert judgment [96]; contingency estimation software using fuzzy arithmetic for addressing the impact of risks on project performance [18]; the application of fuzzy logic for general risk identification and assessment of construction projects [16]; and a critical risk evaluation model using fuzzy-TOPSIS and fuzzy-LINMAP for a gas reservoir construction project [83]. 5. Discussion and future research directions As a basic concept, fuzzy logic/set theory is predominantly used in risk and uncertainty analysis in different sectors of the construction industry [12]. Over the years, the concept has been extended and modified along with many other methods to overcome the drawbacks of basic fuzzy logic/set theory. The study’s finding is that the majority of research applies fuzzy and hybrid methods

126

M.S. Islam et al. / Advanced Engineering Informatics 33 (2017) 112–131

for risk identification, assessment and prioritisation. However, it appears that a comprehensive risk management framework is lacking [12], apart from Salah’s [24] fuzzy set-based overall risk identification, assessment, mitigation, monitoring and control model. Aven et al. [4] have aptly noted some of the limitations of existing risk management tools as the assessment of uncertainties and assignment of probabilities, determination of appropriate values of quantitative parameters, the way of dealing with subjective judgement and objective knowledge, treatment of uncertainties and intangibles, and consequently developed a risk management framework that includes a risk classification structure and procedures for risk decision making. Central to this structure is the classification of possible consequences and associated uncertainties as well as manageability. They argued that the framework is comprehensive enough for classifying the decision situations and characterising and managing risks, but did not propose specific methods for risk analysis and updating, particularly in response to evolving information. Jang et al. [153] proposed a stage-gate risk management framework, that included the source-event and the risk path concept for modelling the causal relationships between risk factors, and the probability-impact-significance of coordination (PIC) method for prioritising critical risks based on manageability and controllability. The ‘‘significance of coordination (C)” factor, which represents the manageability of risk, was incorporated in their model and qualitatively assessed based on expert judgement. They applied the Euclidean distance measuring method instead of sophisticated fuzzy-hybridised methods for computing the aggregated risk score. As an end result, this method provides a relative risk index (RRI) that does not represent the actual level of the project risk, which can misguide decision makers in taking appropriate measures of risk mitigation. Moreover, this method of risk scoring and prioritising is simply an index method, which is unable to capture uncertainties. The method provides a fixed reference point (i.e. crisp number) to define risks, which therefore limits the flexibility of reasonable risk management. To capture the uncertainties and provide flexibility in risk management, some of the studies have developed risk assessment models based on FST [20,21], a basic, practical technique, for the quantitative risk analysis of qualitative data. As an applied form of FST, fuzzy synthetic evaluation has also been frequently used for risk assessment in diverse civil engineering projects, such as building, infrastructure and subway construction [29,31,34]. The advantages of using this method are the simplified form of defining the membership function (i.e., percent frequency of responses) and simple mathematical operations involved in calculating risk scores. A hybrid fuzzy method, e.g., Fuzzy-TOPSIS has been applied in decision making, particularly to select the best option (e.g., contractor, project and risk assessment model). The method is also increasingly popular for the risk assessment of complex projects, such as transport infrastructure, power plant and tunnel construction [25,81,82], because it has an advantage over other methods in capturing additional risk evaluation criteria (e.g., risk detectability and vulnerability) beyond the likelihood and consequence of risks [81]. The FAHP, another hybrid method, is also popular because it is easy to understand and prioritise risks in different hierarchy levels by considering multi-criteria and multi-attributes [57] and was frequently used prior to 2010 for the risk assessment of different types of construction projects [51,53,55,57]. Nonetheless, FAHP is recognised as tedious and inefficient for complex projects having a large number of risks, [77]. As a result, FAHP is now often used in conjunction with other methods, such as fuzzy-TOPSIS, for determining fuzzy linguistic variable weights [8]. These fuzzy based methods of risk assessment have the shortcomings of evaluating the project risk level simply based on conventional aggregation rules for risk rating and ranking, and lack an effective validation process [12,106]. A fuzzy system that relies on fuzzy inference

requires additional development and computational effort such as in membership functions, logical operations or if-then rules, fuzzification and defuzzification tasks, and has problems dealing with judgment-based biases [136]. Previous studies show that subjective judgment-based risk assessment in complex projects with high level of uncertainty, provides biased results, particularly, due to ignorance of causal relationships and dependencies between the variables [66,82,154]. Cox [154] pointed out that causal analysis and interpretation of causal knowledge are often disregarded in risk analysis, thus providing a less reliable outcome for decision making. Risks in complex projects are interdependent and risk dependencies must be realistically treated [94]. However, fuzzy risk assessment models cannot cope with the causal relationships and dependencies between risks [39,155]. Fidan et al. [155] presented an ontology for relating risk-related concepts to cost overruns in international projects, by modelling the causal relations between various risk sources (i.e. risk paths) and sources of vulnerability that interfere with these paths. It should be mentioned that risks dependencies can only be realistically captured by the knowledge and experiences of domain experts. ANP allows the capture of interdependencies between the project risks [82]. That is why researchers have hybridised the fuzzy method with ANP (collectively FANP) and frequently used it for risk management of PPP infrastructure, tunnel and power plant projects [82,92–94]. Nevertheless, the pairwise approach to comparing pairs of risks in FANP is as tedious and lengthy as FAHP, as well as being incapable of updating new information into the risk structure. Thus, when there is an increased number of variables along with their dependencies, complexity increases exponentially in the ANP model. Like ANP, few other methods, for example, Structural Equation Modeling (SEM) [156,157], Bayesian networks [68,70,158,159] and Credal networks [160] also allow the representation of interdependencies between risks. Unlike ANP, the practicality of SEM is not hindered by an increased number of variables [161], but it provides satisfactory results only if the sample size is very large [162–164]. Ironically, eliciting extensive subjective data from experts is very difficult and time consuming, particularly for large, complex construction projects. Risk assessment in the construction industry heavily relies on the previous experience and subjective judgment of the experts, and probability-impact (P-I) models have the potential for modelling complex dependences between the risks [12]. However, the input and output parameters in P-I models are ambiguous, and having only two criteria (i.e. risk probability and impact) limits proper risk assessment [51,153]. Taroun [12] suggested some improvements to existing P-I models, such as adding other parameters reflecting the nature of risks, experience of the experts, and the project environment. The Fuzzy Analytical Network Process (FANP) along with fuzzyTOPSIS can be used to assess the risks of complex projects using different decision criteria, such as the probability, impact, detection and manageability of risks [82], however, computational complexity hinders the viability of this model. As a P-I model, the Bayesian network model has been extensively used in risk analysis due to its superiority in handling both objective and subjective data, model updating capability and reliability testing of outcomes [67,158,165]. Despite of having such positive applications, Bayesian networks require an exact probability of each risk factor elicited from the experts, which is difficult to obtain, particularly in case of lack of knowledge and experience, limited information, time constraints and inability to determine an exact probability [74,160]. Qu and Tang [160] applied a Credal network, which was first introduced by Cozman [166], for developing a risk assessment model for software projects. The model is ideally an extension of the Bayesian network. In a Credal network, a set or a range of

M.S. Islam et al. / Advanced Engineering Informatics 33 (2017) 112–131

possible probabilities instead of a single probability value for a risk event is provided by the experts, and thus the problem of accuracy and completeness of the experts’ knowledge regarding the assessment of uncertainty can be addressed [160,166]. Akin to the Bayesian network, a Credal network can integrate both objective and subjective knowledge in the same model [160,167]. However, updating the model in a Credal network is nonlinear and quite complicated. The method requires an exhaustive algorithm that needs Credal sets to be specified by means of vertices sets. The updating process in a Credal network is not suited to having a large number of vertices in the network [167]. A method is therefore required that can solve the uncertainty, vagueness and imprecision of the risk data, capture the complex relationships of risks and update project progress with new information. Risk is usually assessed by the likelihood of occurrence and consequence of events only, while often disregarding risk detectability and manageability - one of the major drawbacks in the current risk management process – as these aspects need to be considered in risk-assessment methods [4,12,139]. Moreover, vulnerability assessment and management is considered to be an innovative approach to project risk analysis for dealing with project complexity and risk uncertainties [12,168,169]. The level of vulnerability of a project represents its capacity to respond to the risk (i.e. manageability of the risk) and serves an important role in assessing a specific project risk level [4,169]. The Bayesian belief network is capable of handling both risk detectability and vulnerability along with complex relationships between the risks using the probability of the risk events obtained from expert judgments [139], but the BBN requires exact probability data for risk analysis, which limits the application of this method for risk assessment of complex and uncertain construction projects. This limitation needs to be solved by modifying the algorithm of the BBN model or adding other suitable methods, which will allow the model to function with a range or set of probabilities. As an extension of BBN, the Credal network has potential for use in risk and uncertainty modelling, but surprisingly has been disregarded in construction risk management research. Alternatively, a risk assessment method for complex projects developed by the combination of fuzzy logic and BBNs would provide a new dimension for future risk management research. With the fuzzy BBNs method, fuzzy logic first defines and measures the risk evaluation criteria (likelihood of occurrence, level of impact of the risk, risk detectability, mitigation capacity, etc.) according to the verbal opinions of the experts and transfers them into a fuzzy number (set of probabilities instead of a single probability). It defines the fuzzy rules to make the connection between inputs (risk likelihood, consequence, detectability, etc.) and output (risk level). Fuzzy arithmetic is then used to convert the fuzzy outcome into a single value. The BBN is a probability-based network method that is well accepted for handling complex and uncertain relationships between risks. With this method, the level of relationships involved is measured as a single probability value. To capture uncertainty and biasness, fuzzy logic is appropriate for defining the probabilistic relationships as a fuzzy number instead of a single probability value. Novak [104] advised that fuzzy logic based model requires improvement by linking with probability theory, alternatively, Taroun [12] suggested to improve the probability based model by adding other suitable method, so that the modified risk assessment model is able to capture the uncertainties, reduce subjective biasness, and handle causal relationships among the risks. Thus, the FBBN model may ensure the appropriate risk assessment of complex projects. This method can be demonstrated for risk assessment and time or cost analysis by considering the risks for different types of complex projects such as power plants, tunnel construction and other infrastructure projects.

127

While Choi and Mahadevan [17] presented a fuzzy-Bayesian probability model for risk assessment and showed a procedure to update the model with new data, their model does not involve the causal relationships of risks; rather risks are considered to be random by nature. Ren et al. [72] used a fuzzy Bayesian network model for assessing the accident risks of offshore projects, presenting the model on a small scale. However, their omission of any reliability tests of the data incorporated from the experts, who had different levels (i.e. years) of working experience, is a major weakness because risk assessment quality and reliability is directly related to the length of professional experience of the expert [75]. In a recent study, Zhang et al. [74] developed a model combining fuzzy set theory and Bayesian networks for safety risk analysis in tunnel construction. Kabir et al. [75] also presented and applied a FBBN safety assessment model for oil and gas pipeline projects. Since the reliability of the risk data collected from judgments of expert depends on their knowledge, experience, and intuition, reliability analysis of such data becomes necessary. For the reliability of the data set, they presented a modified FBBN model by introducing the level of judgment ability of the expert, which is assigned by clustering the experts’ judgment according to their academic qualifications and work experience. Thus, FBBN is a potential tool that can be extended or applied in a wider range of risk assessment research for other complex construction projects (see Zhang et al. [74] and Karib et al. [75] for more details of FBBN). The FBBN model is not, however without limitations. The process of constructing causal relationships among the risks and its reliability fully depends on the experience and intuitive knowledge of domain experts, and is obviously quite laborious [74]. An ontology of project-specific risk relationships knowledge can greatly improve this process [155]. Moreover, both fuzzy and BBN have limitations in modelling risks and uncertainties with large number of system variables, the most common situation with complex engineering systems [72]. Another drawback of modelling fuzzybased uncertainty is risk ranking, the outcome of risk assessment, which is an important aspect but the most sensitive one. Borgonovo [170] demonstrated different uncertainty measurement methods, such as Pearson correlation coefficient, Spearman correlation coefficient, analysis of variance, and moment-independent (i.e. measuring the importance of uncertainty based on whole output distribution), which are of potential use for discrepancy analysis and subsequently for finding the reliability of the model. Thus, risk assessment by fuzzy and hybrid methods (i.e. FBBN) could be improved in risk ranking and discrepancy testing by applying suitable uncertainty measuring methods as suggested by Borgonovo [170]. In summary, the findings of this review paper of fuzzy and hybrid methods in risk management research partially differ from Chan et al.’s [11] previous review of fuzzy techniques applied in construction management research and critique of available literature published from 1995 to 2005. They concluded that the fuzzy membership function and linguistic variables can be applied to solving the complexity and uncertainty in construction projects. A fuzzy hybrid method, called the fuzzy-artificial neural network (ANN), was claimed to have great potential for modelling uncertainties in complex projects, the assertion being that fuzzy logic is strong in uncertainty modelling, whereas ANN is a viable tool for ‘‘pattern recognition and automatic learning”. However, ANNbased models require a large amount of objective data from similar projects, the critical and most limiting aspect of complex construction projects [74,136]. Most studies acknowledge that construction risk management is highly dependent on subjective expert judgment. Thus, obtaining objective data for complex construction projects for risk assessment is unrealistic, because of the uncertainties and uniqueness of individual construction project [22,54]. Thus, compared with Chan et al.’s [11] findings, our study concludes that

128

M.S. Islam et al. / Advanced Engineering Informatics 33 (2017) 112–131

the FBBN and Credal network have the most potential for future research in overcoming the drawbacks of handling project uncertainties, limited objective data and the dependency on expert judgment for risk assessment.

[5]

[6]

6. Conclusion Fuzzy and hybrid methods have been increasingly used in construction risk-management research and numerous methods have been developed and demonstrated in different construction and engineering management (CEM) fields. Before carrying out further research based on fuzzy and hybrid methods, it is important to understand the basic concepts, benefits and limitations of these methods. This paper provides a comprehensive review of the literature concerning fuzzy-based methods applied in the construction management discipline. The fuzzy papers in this field have been retrieved from the leading CEM journals in the SJR list published between 2005 and 2017. The content analysis revealed three types of fuzzy-based methods to be (1) basic fuzzy (fuzzy logic and FST), (2) extended fuzzy (fuzzy arithmetic, fuzzy synthetic evaluation, fuzzy expert system, fuzzy Mamdani inference, fuzzy comprehensive evaluation and fuzzy consensus qualitative analysis) and (3) hybrid fuzzy method. The latter is broadly classified as fuzzy probability methods, fuzzy matrix methods, fuzzy structured methods, the fuzzy cloud model and fuzzy integral process; where fuzzy probability methods comprise the event tree, fault tree, MCS, Bayesian probability theory, ANN and FMEA; fuzzy matrix methods comprise TOPSIS, VIKOR and COPRAS; and fuzzy structured methods are the AHP, ANP and BBN. Of these methods, FST has been found to be a viable tool for the quantitative risk analysis of qualitative data in establishing a realistic value for project duration or cost. Fuzzy synthetic evaluation has also been frequently and consistently used for risk assessment in a variety of construction projects, such as buildings, infrastructure and subways. The fuzzy-TOPSIS method has been applied consistently in decision making, particularly in selecting from alternatives and the risk assessment of complex projects such as transport infrastructure, power plants and tunnel construction. Risks are interdependent with complex projects and therefore FANP is frequently used in PPP infrastructure, tunnel and power plant projects. The tedious and lengthy calculations needed for the pairwise comparisons involved, however, render this method impractical for risk assessment involving numerous variables, while the model is also incapable of updating new information into the risk structure. To overcome this drawback, F-BBNs is potentially a new dimension for future research because fuzzy logic can capture uncertainty, vagueness and imprecision in risk data, and Bayesian networks can provide an appropriate means of modelling interdependent risks, updating probabilistic information and capturing uncertainty without compromising the reliability of the risk assessment system. In addition, as a modified tool of BBN, Credal networks is also recommended for future research in construction risk management due to their inherent capability of capturing uncertainties in subjective judgments. References [1] W. Mark, P.E. Cohen, R.P. Glen, Project risk identification and management, AACE Int. Trans. INT.01 (2004) 1–5. [2] a. Nieto-Morote, F. Ruz-Vila, A fuzzy approach to construction project risk assessment, Int. J. Proj. Manage. 29 (2011) 220–231, http://dx.doi.org/ 10.1016/j.ijproman.2010.02.002. [3] G. Abdollahzadeh, S. Rastgoo, Risk assessment in bridge construction projects using fault tree and event tree analysis methods based on fuzzy logic, ASCEASME J. Risk Uncert. Eng. Syst., Part B Mech. Eng. 1 (2015) 031006, http://dx. doi.org/10.1115/1.4030779. [4] T. Aven, J.E. Vinnem, H.S. Wiencke, A decision framework for risk management, with application to the offshore oil and gas industry, Reliab.

[7]

[8]

[9]

[10]

[11]

[12]

[13]

[14]

[15]

[16]

[17]

[18]

[19]

[20]

[21]

[22]

[23]

[24] [25]

[26]

[27]

[28]

[29]

[30]

[31]

Eng. Syst. Saf. 92 (2007) 433–448, http://dx.doi.org/10.1016/j. ress.2005.12.009. A. Sohrabinejad, M. Rahimi, Risk determination, prioritization, and classifying in construction project case study: Ghard Tehran commercial-administrative complex, J. Constr. Eng. 2015 (2015). P. Rezakhani, Current state of existing project risk modeling and analysis methods with focus on fuzzy risk assessment – Literature review, Frat. Ed Integrita Strutt. 20 (2012) 17–21, http://dx.doi.org/10.3221/IGF-ESIS.20.02. A.T. Nguyen, L.D. Nguyen, L. Le-hoai, C.N. Dang, Quantifying the complexity of transportation projects using the fuzzy analytic hierarchy process, Int. J. Proj. Manage. (2015), http://dx.doi.org/10.1016/j.ijproman.2015.02.007. O. Taylan, A.O. Bafail, R.M.S. Abdulaal, M.R. Kabli, Construction projects selection and risk assessment by fuzzy AHP and fuzzy TOPSIS methodologies, Appl. Soft Comput. 17 (2014) 105–116, http://dx.doi.org/10.1016/j. asoc.2014.01.003. H. Fares, T. Zayed, Hierarchical fuzzy expert system for risk of failure of water mains, J. Pipeline Syst. Eng. Pract. 1 (2010) 53–62, http://dx.doi.org/10.1061/ (ASCE)PS.1949-1204.0000037. M. Cheng, Y. Lu, Developing a risk assessment method for complex pipe jacking construction projects, Autom. Constr. 58 (2015) 48–59, http://dx.doi. org/10.1016/j.autcon.2015.07.011. A.P.C. Chan, D.W.M. Chan, J.F.Y. Yeung, Overview of the application of ‘‘Fuzzy Techniques” in construction management research, J. Constr. Eng. Manage. 135 (2009) 1241–1252, http://dx.doi.org/10.1061/(ASCE)CO.19437862.0000099. A. Taroun, Towards a better modelling and assessment of construction risk: insights from a literature review, Int. J. Proj. Manage. 32 (2014) 101–115, http://dx.doi.org/10.1016/j.ijproman.2013.03.004. Y. Ke, S. Wang, A.P.C. Chan, E. Cheung, Research trend of public-private partnership in construction journals, J. Constr. Eng. Manage. 135 (2009) 1076–1086. A. Jamshidi, A. Yazdani-Chamzini, S.H. Yakhchali, S. Khaleghi, Developing a new fuzzy inference system for pipeline risk assessment, J. Loss Prev. Process Ind. 26 (2013) 197–208, http://dx.doi.org/10.1016/j.jlp.2012.10.010. S. Kucukali, Risk assessment of river-type hydropower plants using fuzzy logic approach, Energy Policy 39 (2011) 6683–6688, http://dx.doi.org/ 10.1016/j.enpol.2011.06.067. A. Ismail, A.M. Abd, Z. Bin, Chik, Approach to analyze risk factors for construction projects utilizing fuzzy logic.pdf, J. Appl. Sci. 20 (2008) 3738– 3742. H. Choi, S. Mahadevan, Construction project risk assessment using existing database and project-specific information, J. Constr. Eng. Manage. 134 (2008) 894–904. M.M.G. Elbarkouky, N.B. Siraj, A.R. Fayek, Fuzzy Contingency Determinator Ó A Fuzzy Arithmetic-Based Risk Analysis Tool for Construction Projects, IEEE. (2015). Y. Mao, W. Wu, Fuzzy real option evaluation of real estate project based on risk analysis, Syst. Eng. Procedia. 1 (2011) 228–235, http://dx.doi.org/ 10.1016/j.sepro.2011.08.036. P. Pawan, P. Lorterapong, A. Fuzzy-Based, Integrated framework for assessing time contingency in construction projects, J. Constr. Eng. Manage. (2015) 1–9, http://dx.doi.org/10.1061/(ASCE)CO.1943-7862.0001073. A. Maravas, J.P. Pantouvakis, Project cash flow analysis in the presence of uncertainty in activity duration and cost, Int. J. Proj. Manage. 30 (2012) 374– 384, http://dx.doi.org/10.1016/j.ijproman.2011.08.005. M. Malek, D. Ph, A.M. Asce, M. Tumeo, D. Ph, F. Asce, et al., Fuzzy logic approach to risk assessment associated with concrete deterioration, J. Risk Uncertain. Eng. Syst. 1 (2015), http://dx.doi.org/10.1061/AJRUA6.0000811. X. Zhao, B. Hwang, A.M. Asce, S.P. Low, Developing fuzzy enterprise risk management maturity model for construction firms, J. Constr. Eng. Manage. 9 (2013) 1179–1189, http://dx.doi.org/10.1061/(ASCE)CO.1943-7862.0000712. A. Salah, Fuzzy Set-based Risk Management for Construction Projects, Concordia University, 2015. Y. Ji, G.H. Huang, W. Sun, Risk assessment of hydropower stations through an integrated fuzzy entropy-weight multiple criteria decision making method: a case study of the Xiangxi River, Expert Syst. Appl. 42 (2015) 5380–5389, http://dx.doi.org/10.1016/j.eswa.2014.12.026. A.M. Aboushady, M.M. Marzouk, M.M.G. Elbarkouky, Fuzzy consensus qualitative risk analysis framework for building construction projects, IEEE (2013) 1160–1165. K. Mokhtari, J. Ren, C. Roberts, J. Wang, Decision support framework for risk management on sea ports and terminals using fuzzy set theory and evidential reasoning approach, Expert Syst. Appl. 39 (2012) 5087–5103, http://dx.doi. org/10.1016/j.eswa.2011.11.030. Y.-M. Wang, T.M.S. Elhag, A fuzzy group decision making approach for bridge risk assessment, Comput. Ind. Eng. 53 (2007) 137–148, http://dx.doi.org/ 10.1016/j.cie.2007.04.009. H. Abdul-Rahman, C. Wang, Y.L. Lee, Design and pilot run of Fuzzy Synthetic Model (Fsm) for risk evaluation in civil engineering, J. Civ. Eng. Manage. 19 (2013) 217–238, http://dx.doi.org/10.3846/13923730.2012.743926. S. Mu, H. Cheng, M. Chohr, W. Peng, Assessing risk management capability of contractors in subway projects in mainland China, Int. J. Proj. Manage. 32 (2014) 452–460, http://dx.doi.org/10.1016/j.ijproman.2013.08.007. X. Zhao, B.-G. Hwang, Y. Gao, A fuzzy synthetic evaluation approach for risk assessment: a case of Singapore’s Green Projects, J. Clean. Prod. 115 (2015) 203–213, http://dx.doi.org/10.1016/j.jclepro.2015.11.042.

M.S. Islam et al. / Advanced Engineering Informatics 33 (2017) 112–131 [32] E.E. Ameyaw, A.P.C. Chan, O.-M. De-Graft, E. Coleman, A fuzzy model for evaluating risk impacts on variability between contract sum and final account in government-funded construction projects, J. Facil. Manage. 13 (2015) 45–69. [33] Y. Xu, A.P.C. Chan, J.F.Y. Yeung, Developing a fuzzy risk allocation model for ppp projects in China, J. Constr. Eng. Manage. 136 (2010) 884–903, http://dx. doi.org/10.1061/ASCECO.1943-7862.0000189. [34] Y. Xu, J.F.Y. Yeung, A.P.C. Chan, D.W.M. Chan, S.Q. Wang, Y. Ke, Developing a risk assessment model for PPP projects in China-A fuzzy synthetic evaluation approach, Autom. Constr. 19 (2010) 929–943, http://dx.doi.org/10.1016/j. autcon.2010.06.006. [35] R.A. Salawu, F. Abdullah, Assessing risk management maturity of construction organisations on infrastructural project delivery in Nigeria, Procedia - Soc. Behav. Sci. 172 (2015) 643–650, http://dx.doi.org/10.1016/j. sbspro.2015.01.414. [36] A. Idrus, M. Fadhil Nuruddin, M.A. Rohman, Development of project cost contingency estimation model using risk analysis and fuzzy expert system, Expert Syst. Appl. 38 (2011) 1501–1508, http://dx.doi.org/10.1016/j. eswa.2010.07.061. [37] A.R. Fayek, A. Oduba, Predicting industrial construction labor productivity using fuzzy expert systems, J. Constr. Eng. Manage. 131 (2005) 938–941, http://dx.doi.org/10.1061/(ASCE)0733-9364(2005) 131:8(938). [38] A.M. Aboshady, M.M.G. Elbarkouky, M. Mohamed, A Fuzzy Risk Management Framework for the Egyptian Real Estate Development Projects, in: Build. Solut. Archit. Eng. Archit. Eng. Inst. Conf., 2013: pp. 343–352. [39] I. Dikmen, M.T. Birgonul, S. Han, Using fuzzy risk assessment to rate cost overrun risk in international construction projects a, Int. J. Proj. Manage. 25 (2007) 494–505, http://dx.doi.org/10.1016/j.ijproman.2006.12.002. [40] H. Zhao, N. Li, Risk evaluation of a UHV power transmission construction project based on a cloud model and FCE method for sustainability, Sustainability 7 (2015) 2885–2914, http://dx.doi.org/10.3390/su7032885. [41] Q. Yu, S. Huang, J. Du, Fuzzy comprehensive evaluation of risks in subway station construction, in: ICTE 2015, 2015: pp. 2133–2140. [42] Z. Wang, A.M. Asce, L. Qiao, S. Li, L. Bi, Risk assessment for stability and containment property of an underground oil storage facility in construction phase using fuzzy comprehensive evaluation method, ASCE-ASME J. Risk Uncert. Eng. Syst., Part A Civ. Eng. 2 (2016) 1–14, http://dx.doi.org/10.1061/ AJRUA6.0000885. [43] M.-Y. Cheng, N.-D. Hoang, Risk score inference for bridge maintenance project using evolutionary fuzzy least squares support vector machine, J. Comput. Civ. Eng. 28 (2014) 04014003, http://dx.doi.org/10.1061/(ASCE)CP.19435487.0000275. [44] M. Abdelgawad, A.R. Fayek, Fuzzy reliability analyzer: quantitative assessment of risk events in the construction industry using fuzzy faulttree analysis, J. Constr. Eng. Manage. 137 (2011) 294–302, http://dx.doi.org/ 10.1061/(ASCE)CO.1943-7862.0000285. [45] a.V. Thomas, S.N. Kalidindi, L.S. Ganesh, Modelling and assessment of critical risks in BOT road projects, Constr. Manag. Econ. 24 (2006) 407–424, http://dx. doi.org/10.1080/01446190500435275. [46] F. Nasirzadeh, A. Afshar, M. Khanzadi, S. Howick, Integrating system dynamics and fuzzy logic modelling for construction risk management, Constr. Manage. Econ. 26 (2008) 1197–1212, http://dx.doi.org/10.1080/ 01446190802459924. [47] R. Ferdous, F. Khan, R. Sadiq, P. Amyotte, B. Veitch, Fault and event tree analyses for process systems risk analysis: uncertainty handling formulations, Risk Anal. 31 (2011) 86–107, http://dx.doi.org/10.1111/ j.1539-6924.2010.01475.x. [48] L.A. Vidal, F. Marle, J.C. Bocquet, Measuring project complexity using the analytic hierarchy process, Int. J. Proj. Manage. 29 (2011) 718–727, http://dx. doi.org/10.1016/j.ijproman.2010.07.005. [49] J.H. Jung, D.Y. Kim, H.K. Lee, The computer-based contingency estimation through analysis cost overrun risk of public construction project, KSCE J. Civ. Eng. 00 (2015) 1–12, http://dx.doi.org/10.1007/s12205-015-0184-8. [50] a. Mohammadi, M. Tavakolan, Construction project risk assessment using combined fuzzy and FMEA, IEEE (2013) 232–237, http://dx.doi.org/10.1109/ IFSA-NAFIPS.2013.6608405. [51] M. Abdelgawad, A.R. Fayek, Risk management in the construction industry using combined Fuzzy FMEA and Fuzzy AHP, J. Constr. Eng. Manage. 136 (2010) 1028–1036, http://dx.doi.org/10.1061/(ASCE)CO.1943-7862.0000210. [52] M. Abdelgawad, A.R. Fayek, Comprehensive hybrid framework for risk analysis in the construction industry using combined failure mode and effect analysis, fault trees, event trees, and fuzzy logic, J. Constr. Eng. Manage. 138 (2012) 642–651, http://dx.doi.org/10.1061/(ASCE)CO.19437862.0000471. [53] P.X.W. Zou, J. Li, Risk identification and assessment in subway projects: case study of Nanjing Subway Line 2, Constr. Manage. Econ. 28 (2010) 1219–1238, http://dx.doi.org/10.1080/01446193.2010.519781. [54] J. Li, P.X.W. Zou, Fuzzy AHP-based risk assessment methodology for PPP projects, J. Constr. Eng. Manage. 137 (2011) 1205–1209, http://dx.doi.org/ 10.1061/(ASCE)CO.1943-7862.0000362. [55] J. Zeng, M. An, N.J. Smith, Application of a fuzzy based decision making methodology to construction project risk assessment, Int. J. Proj. Manage. 25 (2007) 589–600, http://dx.doi.org/10.1016/j.ijproman.2007.02.006. [56] R.K. Sharma, D. Kumar, P. Kumar, Systematic failure mode effect analysis (FMEA) using fuzzy linguistic modelling, Int. J. Qual. Reliab. Manage. 22 (2005) 986–1004, http://dx.doi.org/10.1108/02656710510625248.

129

[57] P. Chen, J. Wang, Application of a Fuzzy AHP method to risk assessment of international construction projects, 2009 Int. Conf. Electron. Commer. Bus. Intell. (2009) 459–462. doi:10.1109/ECBI.2009.14. [58] H. Subramanyan, P.H. Sawant, V. Bhatt, Construction project risk assessment: development of model based on investigation of opinion of construction project experts from India, J. Constr. Eng. Manage. 138 (2012) 409–421, http://dx.doi.org/10.1061/(ASCE)CO.1943-7862. [59] G. Zhang, P. Zou, Fuzzy analytical hierarchy process risk assessment approach for joint venture construction projects in China, J. Constr. Eng. Manage. 133 (2007) 771–779, http://dx.doi.org/10.1061/(ASCE)0733-9364(2007) 133:10 (771). [60] R. Sonmez, Range estimation of construction costs using neural networks with bootstrap prediction intervals, Expert Syst. Appl. 38 (2011) 9913–9917, http://dx.doi.org/10.1016/j.eswa.2011.02.042. [61] C. Koo, T. Hong, C. Hyun, K. Koo, A CBR-based hybrid model for predicting a construction duration and cost based on project characteristics in multifamily housing projects, Can. J. Civ. Eng. 37 (2010) 739–752, http://dx.doi.org/ 10.1139/L10-007. [62] M.-Y. Cheng, H.-C. Tsai, W.-S. Hsieh, Web-based conceptual cost estimates for construction projects using evolutionary fuzzy neural inference model, Autom. Constr. 18 (2009) 164–172, http://dx.doi.org/10.1016/j.autcon.2008.07.001. [63] Y. Wang, T. Elhag, An adaptive neuro-fuzzy inference system for bridge risk assessment, Expert Syst. Appl. 34 (2008) 3099–3106, http://dx.doi.org/ 10.1016/j.eswa.2007.06.026. [64] W. Imbeah, S. Guikema, Managing construction projects using the advanced programmatic risk analysis and management model, J. Constr. Eng. Manage. 135 (2009) 772–781, http://dx.doi.org/10.1061/(asce)0733-9364(2009) 135:8(772). [65] M. Farajian, Q. Cui, A. Multi-Objective, Decision support system for PPP funding decisions, Comput. Civ. Eng. (2011) 760–767, http://dx.doi.org/ 10.1061/41182(416)94. [66] V. Khodakarami, A. Abdi, Project cost risk analysis: a Bayesian networks approach for modeling dependencies between cost items, Int. J. Proj. Manage. 32 (2014) 1233–1245, http://dx.doi.org/10.1016/j.ijproman.2014.01.001. [67] S. Tesfamariam, Risk assessment of bridges using Bayesian belief network : with consideration of seismic and aging, 2013. [68] N. Khakzad, F. Khan, P. Amyotte, Quantitative risk analysis of offshore drilling operations: a Bayesian approach, Saf. Sci. 57 (2013) 108–117, http://dx.doi. org/10.1016/j.ssci.2013.01.022. [69] G. Medina-Oliva, P. Weber, C. Simon, B. Iung, Bayesian networks applications on dependability, risk analysis and maintenance, IFAC Proc. 2 (2009) 215– 220, http://dx.doi.org/10.3182/20090610-3-IT-4004.00042. [70] P. Weber, G. Medina-Oliva, C. Simon, B. Iung, Overview on Bayesian networks applications for dependability, risk analysis and maintenance areas, Eng. Appl. Artif. Intell. 25 (2012) 671–682, http://dx.doi.org/10.1016/j. engappai.2010.06.002. [71] I.C. Cárdenas, S.S.H. Al-Jibouri, J.I.M. Halman, F.a. Van Tol, Modeling riskrelated knowledge in tunneling projects, Risk Anal. 34 (2014) 323–339, http://dx.doi.org/10.1111/risa.12094. [72] J. Ren, I. Jenkinson, J. Wang, D.L. Xu, J.B. Yang, An offshore risk analysis method using fuzzy bayesian network, J. Offshore Mech. Arct. Eng. 131 (2009) 1–28, http://dx.doi.org/10.1115/1.3124123. [73] P.C. Li, G.H. Chen, L.C. Dai, L. Zhang, A fuzzy bayesian network approach to improve the quantification of organizational influences in HRA frameworks, Saf. Sci. 50 (2012) 1569–1583, http://dx.doi.org/10.1016/j.ssci.2012.03.017. [74] L. Zhang, X. Wu, Y. Qin, M.J. Skibniewski, W. Liu, Towards a fuzzy bayesian network based approach for safety risk analysis of tunnel-induced pipeline damage, Risk Anal. 36 (2016) 278–301, http://dx.doi.org/10.1111/risa.12448. [75] G. Kabir, R. Sadiq, S. Tesfamariam, A fuzzy Bayesian belief network for safety assessment of oil and gas pipelines, Struct. Infrastruct. Eng. 2479 (2016) 1–16, http://dx.doi.org/10.1080/15732479.2015.1053093. [76] Z. Lin, Y. Jianping, Risk Assessment based On Fuzzy Network (F-ANP) in new campus construction project, Syst. Eng. Procedia. 1 (2011) 162–168, http:// dx.doi.org/10.1016/j.sepro.2011.08.027. [77] S. Ebrahimnejad, S.M. Mousavi, R. Tavakkoli-Moghaddam, H. Hashemi, B. Vahdani, A novel two-phase group decision making approach for construction project selection in a fuzzy environment, Appl. Math. Model. 36 (2012) 4197–4217, http://dx.doi.org/10.1016/j.apm.2011.11.050. [78] M. Ebrat, R. Ghodsi, Construction project risk assessment by using adaptivenetwork-based fuzzy inference system: an empirical study, KSCE J. Civ. Eng. 18 (2014) 1213–1227, http://dx.doi.org/10.1007/s12205-014-0139-5. [79] J. Tamosaitience, E.K. Zavadskas, Z. Turskis, Multi-criteria risk assessment of a construction project, Procedia Comput. Sci. 17 (2013) 129–133, http://dx.doi. org/10.1016/j.procs.2013.05.018. [80] A. Karimiazari, N. Mousavi, S.F. Mousavi, S. Hosseini, Risk assessment model selection in construction industry, Expert Syst. Appl. 38 (2011) 9105–9111, http://dx.doi.org/10.1016/j.eswa.2010.12.110. [81] M.M. Fouladgar, A. Yazdani-Chamzini, E.K. Zavadskas, Risk evaluation of tunneling projects, Arch. Civ. Mech. Eng. 12 (2012) 1–12, http://dx.doi.org/ 10.1016/j.acme.2012.03.008. [82] S.H. Zegordi, E. Rezaee Nik, A. Nazari, Power plant project risk assessment using a fuzzy-ANP and fuzzy-TOPSIS method, Int. J. Eng. Trans. B Appl. 25 (2012) 107–120, http://dx.doi.org/10.5829/idosi.ije.2012.25.02b.04. [83] S. Ebrahimnejad, S.M. Mousavi, S.M.H. Mojtahedi, A model for risk evaluation in construction projects based on fuzzy MADM, 2008 4th IEEE Int. Conf. Manag. Innov. Technol. (2008) 305–310. doi:10.1109/ICMIT.2008.4654381.

130

M.S. Islam et al. / Advanced Engineering Informatics 33 (2017) 112–131

[84] S. Ebrahimnejad, S.M. Mousavi, H. Seyrafianpour, Risk identification and assessment for build–operate–transfer projects: a fuzzy multi attribute decision making model, Expert Syst. Appl. 37 (2010) 575–586, http://dx.doi. org/10.1016/j.eswa.2009.05.037. [85] M. Yazdani, A. Alidoosti, E.K. Zavadskas, Risk analysis of critical infrastructures using fuzzy copras, Ekon. Istraz. 24 (2011) 27–40, http://dx. doi.org/10.1080/1331677X.2011.11517478. [86] Y.-C. Kuo, S.-T. Lu, Using fuzzy multiple criteria decision making approach to enhance risk assessment for metropolitan construction projects, Int. J. Proj. Manage. 31 (2013) 602–614, http://dx.doi.org/10.1016/j. ijproman.2012.10.003. [87] O. Duran, N. Rodriguez, L.A. Consalter, Neural networks for cost estimation of shell and tube heat exchangers, Expert Syst. Appl. 36 (2009) 7435–7440, http://dx.doi.org/10.1016/j.eswa.2008.09.014. [88] M.-Y. Cheng, H.-C. Tsai, E. Sudjono, Conceptual cost estimates using evolutionary fuzzy hybrid neural network for projects in construction industry, Expert Syst. Appl. 37 (2010) 4224–4231, http://dx.doi.org/ 10.1016/j.eswa.2009.11.080. [89] X. Jin, Neurofuzzy decision support system for efficient risk allocation in public-private partnership infrastructure projects, J. Comput. Civ. Eng. 24 (2010) 525–539, http://dx.doi.org/10.1061/(ASCE)CP.19435487.0000058. [90] F.-Y. Hsiao, S.-H. Wang, W.-C. Wang, C.-P. Wen, W.-D. Yu, Neuro-fuzzy cost estimation model enhanced by fast messy genetic algorithms for semiconductor hookup construction, Comput. Civ. Infrastruct. Eng. 27 (2012) 764–781, http://dx.doi.org/10.1111/j.1467-8667.2012.00786.x. [91] A.S. Bu-Qammaz, I. Dikmen, M.T. Birgonul, Risk assessment of international construction projects using the analytic network process, Can. J. Civ. Eng. 36 (2009) 1170–1181, http://dx.doi.org/10.1139/L09-061. [92] M. Shafiee, A fuzzy analytic network process model to mitigate the risks associated with offshore wind farms, Expert Syst. Appl. 42 (2015) 2143–2152, http://dx.doi.org/10.1016/j.eswa.2014.10.019. [93] A. Valipour, N. Yahaya, N. Md Noor, S. Kildiene˙, H. Sarvari, A. Mardani, A fuzzy analytic network process method for risk prioritization in freeway PPP projects: an Iranian case study, J. Civ. Eng. Manage. 21 (2015) 933–947, http://dx.doi.org/10.3846/13923730.2015.1051104. [94] W. Yan, L. Baoguo, Q. Yi, Fuzzy analytic network process and its application in tunnel engineering risk analysis, EJGE 20 (2015) 6685–6701. [95] W.-C. Wang, R.-J. Dzeng, Y.-H. Lu, Integration of simulation-based cost model and multi-criteria evaluation model for bid price decisions, Comput. Civ. Infrastruct. Eng. 22 (2007) 223–235, http://dx.doi.org/10.1111/j.14678667.2007.00480.x. [96] A.a. Shaheen, A.R. Fayek, S.M. AbouRizk, Fuzzy numbers in cost range estimating, J. Constr. Eng. Manag. 133 (2007) 325–334, doi:10.1061/(ASCE) 0733-9364(2007) 133:4(325). [97] B. Möller, M. Beer, Engineering computation under uncertainty - capabilities of non-traditional models, Comput. Struct. 86 (2008) 1024–1041, http://dx. doi.org/10.1016/j.compstruc.2007.05.041. [98] N. Sadeghi, a.R. Fayek, W. Pedrycz, Fuzzy Monte Carlo simulation and risk assessment in construction, Comput. Civ. Infrastruct. Eng. 25 (2010) 238–252, http://dx.doi.org/10.1111/j.1467-8667.2009.00632.x. [99] G. Heravi, S. Faeghi, Group decision making for stochastic optimization of time, cost, and quality in construction projects, J. Comput. Civ. Eng. 28 (2012) 275–283, http://dx.doi.org/10.1061/(ASCE)CP.1943-5487.0000264. [100] M. Attarzadeh, D. Ph, D.K.H. Chua, D. Ph, M. Asce, M. Beer, et al., Fuzzy randomness simulation of long-term infrastructure projects, ASCE-ASME J. Risk Uncert. Eng. Syst., Part A Civ. Eng. (2017) 1–15, http://dx.doi.org/ 10.1061/AJRUA6.0000902. [101] N. Sadeghi, A. Robinson Fayek, N. Gerami Seresht, Queue performance measures in construction simulation models containing subjective uncertainty, Autom. Constr. 60 (2015) 1–11, http://dx.doi.org/10.1016/j. autcon.2015.07.023. [102] A.M. Abd, Building projects risks sources: conceptual fuzzy assessment approach, Int. J. Eng. Sci. Res. Technol. 4 (2015) 325–332. [103] T. Lyons, M. Skitmore, Project risk management in the Queensland engineering construction industry: a survey, Int. J. Proj. Manage. 22 (2004) 51–61. [104] V. Novak, Reasoning about mathematical fuzzy logic and its future, Fuzzy Sets Syst. 192 (2012) 25–44, http://dx.doi.org/10.1016/j.fss.2010.09.019. [105] V. Novák, Fuzzy sets as a special mathematical model of vagueness phenomenon, Comput. Intell. Theory Appl. Int. Conf. 9th Fuzzy Days Dortmund, Ger. Sept. 18–20, 2006 Proc. (2006) 683–690. doi:10.1007/3540-34783-6_66. [106] D. Dubois, The role of fuzzy sets in decision sciences: old techniques and new directions, Fuzzy Sets Syst. 184 (2011) 3–28, http://dx.doi.org/10.1016/j. fss.2011.06.003. [107] R. Kangari, Construction risk management, Civ. Eng. Syst. 5 (1988) 114–120, http://dx.doi.org/10.1080/02630258808970516. [108] D. Baloi, A.D.F. Price, Modelling global risk factors affecting construction cost performance, Int. J. Proj. Manage. 21 (2003) 261–269, http://dx.doi.org/ 10.1016/S0263-7863(02)00017-0. [109] R. Kangari, L.S. Riggs, Construction risk assessment by linguistics, IEEE Trans. Eng. Manage. 36 (1989) 126–131, http://dx.doi.org/10.1109/17.18829. [110] H.N. Cho, H.H. Choi, Y.B. Kim, A risk assessment methodology for incorporating uncertainties using fuzzy concepts, Reliab. Eng. Syst. Saf. 78 (2002) 173–183, http://dx.doi.org/10.1016/S0951-8320(02)00158-8.

[111] M. Hanss, Applied Fuzzy Arithmetic, Springer, Berlin Heidelberg, 2005, http:// dx.doi.org/10.1017/CBO9781107415324.004, New York. [112] Z. Xu, K. Gao, T.M. Khoshgoftaar, N. Seliya, System regression test planning with a fuzzy expert system, Inf. Sci. (Ny) 259 (2014) 532–543, http://dx.doi. org/10.1016/j.ins.2010.09.012. [113] L.A. Zadeh, Is there a need for fuzzy logic?, Inf Sci. (Ny) 178 (2008) 2751– 2779, http://dx.doi.org/10.1016/j.ins.2008.02.012. [114] Y. Jin, Advanced Fuzzy Systems Design and Applications, Springer, New York, 2003, http://dx.doi.org/10.1007/978-3-7908-1771-3. [115] M. Fasanghari, G.A. Montazer, Design and implementation of fuzzy expert system for Tehran stock exchange portfolio recommendation, Expert Syst. Appl. 37 (2010) 6138–6147, http://dx.doi.org/10.1016/j.eswa.2010.02.114. [116] L. Chen, L. Ma, Z. Liang, An evaluation computing method based on cloud model with core space and its application: bridges management evaluation, Emerg. Technol. Inf. Syst. Comput. Manage. (2013) 1289, http://dx.doi.org/ 10.1007/978-1-4614-7010-6. [117] M. Goldstein, Subjective bayesian analysis : principles and practice, Bayesian Anal. 1 (2006) 403–420. [118] D. Guyonnet, B. Bourgine, D. Dubois, H. Fargier, B. Colme, J.-P. Chilès, Hybrid approach for addressing uncertainty in risk assessments, J. Environ. Eng. 129 (2003) 68–78, http://dx.doi.org/10.1061/(ASCE)0733-9372(2003) 129:1(68). [119] P. Loizou, N. French, Risk and uncertainty in development: a critical evaluation of using the Monte Carlo simulation method as a decision tool in real estate development projects, J. Prop. Invest. Financ. 30 (2012) 198– 210, http://dx.doi.org/10.1016/j.autcon.2010.02.008. [120] M. Attalla, T. Hegazy, Predicting cost deviation in reconstruction projects: artificial neural networks versus regression, J. Constr. Eng. Manage. 129 (2003) 405–411, http://dx.doi.org/10.1061/(ASCE)0733-9364(2003) 129:4(405). [121] G.H. Kim, D.S. Seo, K.I. Kang, Hybrid models of neural networks and genetic algorithms for predicting preliminary cost estimates, J. Comput. Civ. Eng. 19 (2005) 208–211, http://dx.doi.org/10.1061/(ASCE)0887-3801(2005) 19:2 (208). [122] D.P. Alex, M. Al Hussein, A. Bouferguene, S. Fernando, Artificial neural network model for cost estimation: city of Edmonton’s water and sewer installation services, J. Constr. Eng. Manage. 136 (2010) 745–756, http://dx. doi.org/10.1061/(ASCE)CO.1943-7862.0000184. [123] O. Duran, J. MacIel, N. Rodriguez, Comparisons between two types of neural networks for manufacturing cost estimation of piping elements, Expert Syst. Appl. 39 (2012) 7788–7795, http://dx.doi.org/10.1016/j.eswa.2012.01.095. [124] H.-C. Liu, L. Liu, N. Liu, L.-X. Mao, Risk evaluation in failure mode and effects analysis with extended VIKOR method under fuzzy environment, Expert Syst. Appl. 39 (2012) 12926–12934, http://dx.doi.org/10.1016/j.eswa.2012.05.031. [125] K. Xu, L.C. Tang, M. Xie, S.L. Ho, M.L. Zhu, Fuzzy assessment of FMEA for engine systems, Reliab. Eng. Syst. Saf. 75 (2002) 17–29, http://dx.doi.org/ 10.1016/S0951-8320(01)00101-6. [126] T.C. Wang, T.H. Chang, Application of TOPSIS in evaluating initial training aircraft under a fuzzy environment, Expert Syst. Appl. 33 (2007) 870–880, http://dx.doi.org/10.1016/j.eswa.2006.07.003. [127] S. Opricovic, G.H. Tzeng, Compromise solution by MCDM methods: a comparative analysis of VIKOR and TOPSIS, Eur. J. Oper. Res. 156 (2004) 445–455, http://dx.doi.org/10.1016/S0377-2217(03)00020-1. [128] W.H. Tsai, W.C. Chou, J. Der Leu, An effectiveness evaluation model for the web-based marketing of the airline industry, Expert Syst. Appl. 38 (2011) 15499–15516, http://dx.doi.org/10.1016/j.eswa.2011.06.009. [129] K. Devi, Extension of VIKOR method in intuitionistic fuzzy environment for robot selection, Expert Syst. Appl. 38 (2011) 14163–14168, http://dx.doi.org/ 10.1016/j.eswa.2011.04.227. [130] S. Opricovic, Fuzzy VIKOR with an application to water resources planning, Expert Syst. Appl. 38 (2011) 12983–12990, http://dx.doi.org/10.1016/j. eswa.2011.04.097. [131] D. Jato-Espino, E. Castillo-Lopez, J. Rodriguez-Hernandez, J.C. CanterasJordana, A review of application of multi-criteria decision making methods in construction, Autom. Constr. 45 (2014) 151–162, http://dx.doi.org/ 10.1016/j.autcon.2014.05.013. [132] M. Azizi, M. Modarres, A strategic model for location selection of wood industry: an application of ANP, J. Appl. Sci. 7 (2007) 326–333. [133] J. Cheng, R. Greiner, J. Kelly, D. Bell, W. Liu, Learning Bayesian networks from data: an information-theory based approach, Artif. Intell. 137 (2002) 43–90, http://dx.doi.org/10.1016/S0004-3702(02)00191-1. [134] D. Heckerman, Bayesian networks for data mining, Data Min. Knowl. Discov. 1 (1997) 79–119. [135] U.B. Kjærulff, A.L. Madsen, Bayesian Networks and Influence Diagrams, 2008, http://dx.doi.org/10.1007/978-0-387-74101-7. [136] I.C. Cárdenas, S.S.H. Al-Jibouri, J.I.M. Halman, F.a. van Tol, Capturing and integrating knowledge for managing risks in tunnel works, Risk Anal. 33 (2013) 92–108, http://dx.doi.org/10.1111/j.1539-6924.2012.01829.x. [137] J.H. Sigurdsson, L.a. Walls, J.L. Quigley, Bayesian belief nets for managing expert judgement and modelling reliability, Qual. Reliab. Eng. Int. 17 (2001) 181–190, http://dx.doi.org/10.1002/qre.410. [138] J.T. Christian, Geotechnical engineering reliability: how well do we know what we are doing?, J Geotech. Geoenviron. Eng. 130 (2004) 985–1003, http://dx.doi.org/10.1061/(ASCE)1090-0241(2004) 130:10(985). [139] M. Sättele, M. Bründl, D. Straub, Reliability and effectiveness of early warning systems for natural hazards: concept and application to debris flow warning, Reliab. Eng. Syst. Saf. 142 (2015) 192–202, http://dx.doi.org/10.5194/nhessd3-4479-2015.

M.S. Islam et al. / Advanced Engineering Informatics 33 (2017) 112–131 [140] C. Ji, T. Hong, C. Hyun, CBR revision model for improving cost prediction accuracy in multifamily housing projects, J. Manage. Eng. 26 (2010) 229–236, http://dx.doi.org/10.1061/(ASCE)ME.1943-5479.0000018. [141] A.G. Eleye-Datubo, A. Wall, J. Wang, Marine and offshore safety assessment by incorporative risk modeling in a fuzzy-Bayesian network of an induced mass assignment paradigm, Risk Anal. 28 (2008) 95–112, http://dx.doi.org/ 10.1111/j.1539-6924.2008.01004.x. [142] P. cheng Li, G. hua Chen, L. cao Dai, L. Zhang, A fuzzy Bayesian network approach to improve the quantification of organizational influences in HRA frameworks, Saf. Sci. 50 (2012) 1569–1583, http://dx.doi.org/10.1016/j. ssci.2012.03.017. [143] Y.W. Chen, G.H. Tzeng, Using fuzzy integral for evaluating subjectively perceived travel costs in a traffic assignment model, Eur. J. Oper. Res. 130 (2001) 653–664, http://dx.doi.org/10.1016/S0377-2217(99)00439-7. [144] A. Eriksson, Fuzzy risk assessment for construction projects - with focus on farm-based grain drying facilities, Swedish University of Agricultural Sciences, 2015. [145] M.S. Islam, B. Trigunarsyah, M. Hassanain, S. Assaf, Causes of delay in construction projects in Bangladesh, in: 6th Int. Conf. Constr. Eng. Proj. Manag. (ICCEPM 2015), 2015: pp. 82–86. http://www.scopus.com/inward/ record.url?eid=2-s2.0-84927588873&partnerID=40&md5= 0bd34bb52feeffe197bf70b90416046a. [146] F. Mafakheri, M. Breton, S. Chauhan, Project-to-organization matching, Int. J. Inf. Technol. Proj. Manage. 3 (2012) 45–59, http://dx.doi.org/10.4018/ jitpm.2012070104. [147] I.C. Cárdenas, S.S.H. Al-Jibouri, J.I.M. Halman, W. van de Linde, F. Kaalberg, Using prior risk-related knowledge to support risk management decisions: lessons learnt from a tunneling project, Risk Anal. 34 (2014) 1923–1943, http://dx.doi.org/10.1111/risa.12213. [148] O. Špacˇková, J. Šejnoha, D. Straub, Probabilistic assessment of tunnel construction performance based on data, Tunn. Undergr. Sp. Technol. 37 (2013) 62–78, http://dx.doi.org/10.1016/j.tust.2013.02.006. [149] D. Straub, Reliability updating with equality information, Probabilist. Eng. Mech. 26 (2011) 254–258, http://dx.doi.org/10.1016/j.probengmech.2010. 08.003. [150] I. Papaioannou, D. Straub, Reliability updating in geotechnical engineering including spatial variability of soil, Comput. Geotech. 42 (2012) 44–51, http://dx.doi.org/10.1016/j.compgeo.2011.12.004. [151] C. Camós, C. Molins, O. Špacˇková, D. Straub, Development of a probabilistic model for the prediction of building damage due to tunneling induced settlements, in: Geotech. Saf. Risk IV - Proc. 4th Int. Symp. Geotech. Saf. Risk, ISGSR 2013, 2014: pp. 485–491. http://www.scopus.com/inward/record.url? eid=2-s2.0-84896867491&partnerID=tZOtx3y1. [152] C. Camós, O. Špacˇková, D. Straub, C. Molins, Probabilistic approach to assessing and monitoring settlements caused by tunneling, Tunn. Undergr. Sp. Technol. 51 (2016) 313–325, http://dx.doi.org/10.1016/j.tust.2015.10.041. [153] W. Jang, S.H. Han, H. Park, A stage-gate integrated risk control system for LNG plant projects: focusing on the design phase, J. Nat. Gas Sci. Eng. 22 (2015) 437–446, http://dx.doi.org/10.1016/j.jngse.2014.12.018. [154] L.A. Cox, Improving causal inferences in risk analysis, Risk Anal. 33 (2013) 1762–1771, http://dx.doi.org/10.1111/risa.12072.

131

[155] G. Fidan, I. Dikmen, a.M. Tanyer, M.T. Birgonul, Ontology for relating risk and vulnerability to cost overrun in international projects, J. Comput. Civ. Eng. 25 (2011) 302–315, http://dx.doi.org/10.1061/(ASCE)CP.1943-5487.0000090. [156] M. Eybpoosh, I. Dikmen, M.T. Birgonul, Identification of risk paths in international construction projects using structural equation modeling, J. Constr. Eng. Manage. 137 (2011) 1164–1175, http://dx.doi.org/10.1061/ (ASCE)CO.1943-7862.0000382. [157] A. Hameed Memon, I. Abdul Rahman, A.A. Abdul Aziz, N.H. Abdullah, Using structural equation modelling to assess effects of construction resource related factors on cost overrun, World Appl. Sci. J. 21 (2013) 6–15, http://dx. doi.org/10.5829/idosi.wasj.2013.21.mae.9995. [158] L. Podofillini, L. Mkrtchyan, V.N. Dang, Quantification of Bayesian Belief Net Relationships for HRA from Operational Event Analyses, in: Probabilistic Saf. Assess. Manag. PSAM 12, June 2014, Honolulu, Hawaii, 2014. [159] M.R. Martins, M.C. Maturana, Application of Bayesian Belief networks to the human reliability analysis of an oil tanker operation focusing on collision accidents, Reliab. Eng. Syst. Saf. 110 (2013) 89–109, http://dx.doi.org/ 10.1016/j.ress.2012.09.008. [160] Y. Qu, X.-L. Tang, Software project risk assessing model based on credal networks, 2010 Int. Conf. Mach. Learn. Cybern. (2010) 1976–1979. doi:10.1109/ICMLC.2010.5580516. [161] R.P. Bagozzi, Y. Yi, Specification, evaluation, and interpretation of structural equation models, J. Acad. Mark. Sci. 40 (2012) 8–34, http://dx.doi.org/ 10.1007/s11747-011-0278-x. [162] C. Fornell, F.L. Bookstein, Two structural equation models: LISREL and PLS applied to consumer exit-voice theory, J. Mark. Res. 19 (1982) 440, http://dx. doi.org/10.2307/3151718. [163] W.W. Chin, E. Marcolin, P.R. Newsted, A partial least squares latent variable modeling approach for measuring interaction effects: results from a Monte Carlo simulation study and an electronic mail emotion/adoption study, Inf. Syst. Res. 14 (2003) 189–217. [164] P.B. Lowry, J. Gaskin, Partial least squares (PLS) structural equation modeling (SEM) for building and testing behavioral causal theory: when to choose it and how to use it, IEEE Trans. Prof. Commun. 57 (2014) 123–146, http://dx. doi.org/10.1109/TPC.2014.2312452. [165] A.G. Wilson, A.V. Huzurbazar, Bayesian networks for multilevel system reliability, Reliab. Eng. Syst. Saf. 92 (2007) 1413–1420, http://dx.doi.org/ 10.1016/j.ress.2006.09.003. [166] F.G. Cozman, Credal networks, Artif. Intell. 120 (2000) 199–233, http://dx.doi. org/10.1016/S0004-3702(00)00029-1. [167] A. Antonucci, A. Salvetti, M. Zaffalon, Credal Networks for Hazard Assessment of Debris Flows, 2007. [168] L.-A. Vidal, F. Marle, A systems thinking approach for project vulnerability management, Kybernetes 41 (2012) 206–228, http://dx.doi.org/10.1108/ 03684921211213043. [169] H. Zhang, A redefinition of the project risk process: using vulnerability to open up the event-consequence link, Int. J. Proj. Manage. 25 (2007) 694–701, http://dx.doi.org/10.1016/j.ijproman.2007.02.004. [170] E. Borgonovo, Measuring uncertainty importance: investigation and comparison of alternative approaches, Risk Anal. 26 (2006) 1349–1361, http://dx.doi.org/10.1111/j.1539-6924.2006.00806.x.