Dynamic operations and manpower scheduling for ...

SIMTech technical reports (STR_V10_N3_05_POM)

Volume 10 Number 3 Jul-Sep 2009

Dynamic operations and manpower scheduling for high-mix, low-volume precision engineering manufacturing T. J. Chua, T. X. Cai, N. B. Ho, J. M. W. Low, and L. S. Chai

Abstract – The shop floor of a Precision Engineering (PE) company is a typical high-mix low-volume job shop environment in which jobs with variable routings are being processed by machines capable of performing multiple operations. In most scenarios, it is also staffed with skilled craftsman with different job grades and working on different shift patterns. In addition, the production floor is always subjected to ad-hoc and last minute urgent customer orders. This paper aims to explore and analyze a new heuristics and optimization algorithm to solve the dynamic operations and manpower scheduling problems in this complex environment. With this system, companies would be able to improve their delivery performance through effective handling of ad-hoc customer orders and the ability to react to deviations or unplanned events in production.

pairing a schedule that has been disrupted; 2) methods for creating a schedule that is robust with respect to disruptions; and 3) studies of how rescheduling policies affect the performance of the dynamic manufacturing system. Our research focus is on the Rescheduling Methods which describes how schedules are generated and updated. The scheduling of manpower to meet the production schedule requirements forms another important aspect of this project. In the manufacturing context, staff works in shifts. The shift scheduling problem involves determining the number of employees to be assigned to various shifts and the timing of their breaks within the limits allowed by company requirements. In this study, we examined the design and development of the heuristics and optimisation algorithms for effective solving of the operations and manpower scheduling problems. A prototype dynamic operations and manpower scheduling was developed to demonstrate the feasibility of the developed algorithms in solving the scheduling problems involving operations and manpower.

Keywords: Dynamic operations, Manpower scheduling, High-mix, Low volume

1

BACKGROUND

3

In Singapore, Precision Engineering (PE) industry is the backbone of the manufacturing sector; it supports a large number of manufacturing industries such as Electronics, Chemicals, Pharmaceuticals, Biotechnology, Aerospace, Oil & Gas and Automotive. From our recent visits to the companies in the PE industry, it was observed that given such a dynamic and stochastic manufacturing environment, most of these PE companies are still predominantly adopting Microsoft Excel as their planning & scheduling tool. However, based our recent company visits, there is an increasing demand for a computerized production scheduling tool. The common problems highlighted include inability to provide realistic delivery commitment, machine and manpower resources are not optimally utilized to fulfill customer orders, and unacceptable order cycle time. 2

METHODOLOGY

The research approach is as follow: The existing rescheduling methodology and manpower scheduling approaches were surveyed and investigated. Based on the finding and the actual requirements from the companies in the PE industry, new operations rescheduling and manpower scheduling algorithms are proposed and developed. The new algorithms and the interactive gantt chart user interface are then integrated with the job shop scheduling engine. The current job shop engine needs further enhancements in order to cater for the requirements of the PE companies. Extensive test and verification are performed for each set of algorithms before the integration to form the prototype system for dynamic operations and manpower scheduling. Upon the completion of prototype system, it is subjected to test and debugging using realistic data from a PE company.

OBJECTIVE

3.1

The objectives of this project is to explore and develop new heuristics and optimisation algorithms to facilitate handling of the dynamic operations rescheduling and manpower scheduling operational scenarios in a high-mix, low-volume production environment. In the literature on rescheduling, there are three primary types of studies: 1) methods for re-

Literature Review

In the management of dynamic production operations, rescheduling is practically mandatory in order to cater for the unexpected events and disruptions. There are many types of disturbances that can upset the plan, including machine failures, processing time delays, rush orders, quality problems, and unavailable materials.

153

Dynamic operations and manpower scheduling for high-mix, low-volume precision engineering manufacturing

As Bean et al (1991) state, rescheduling is a dynamic approach that responds to disruptions, yet it considers future information (by creating a plan for the future). Guilherme et al (2003) proposed that there are three primary types of studies in the literature of rescheduling: one, methods for repairing a schedule that has been disrupted; two, methods for creating a schedule that is robust with respect to disruptions; and three, studies of how rescheduling policies affect the performance of the dynamic manufacturing system. The scheduling of manpower to meet the production plan requirements forms another important aspect of this project. In the manufacturing context, staff works in shifts. The shift scheduling problem, as defined by Aykin (2006), involves determining the number of employees to be assigned to various shifts and the timing of their breaks within the limits allowed by legal, union, and company requirements. Grabot and Letouzey (2000) noted that the workforce allocation has seldom been addressed with its specificity in the research literature on manufacturing. At the shop-floor level, the specificity of the human resource has been poorly taken into account in scheduling, except in some specific works like Burns and Carter (1985), Emmons (1985) or Hung (1994) all on the case of a regular demand covered by cyclic shift patterns. Yano and Rachamadugu (1991) discussed the use of utility workers in the overloaded parts of the assembly line. Russell, Huang, and Leu (1991) simulated the performance of different labor scheduling policies and labor allocation strategies in a manufacturing cell. Lagodimos and Leopoulos (2000) proposed an algorithm to decide how many unskilled workers to employ in every daily work-shift of some particular month of the master schedule so as to complete the production load of each packing line, while maximizing workforce utilization.

variable product routing. The manpower scheduling module is incorporated to generate the optimised staffing plan to support the production schedule. The generated schedule is presented in an interactive gantt chart to cater for manual adjustment due to ad-hoc and last minute urgent customer orders. These combinations of production scheduling features fill the technology gap of the known commercially available production scheduling software.

3.2

3.2.1

Table 1. Operational characteristics of PE companies. SMEs in Precision Engineering Variable routing Machines capable of performing different operations (Flexible equipment) Process layout Low volume, High Mix Capacity is difficult to define Labour intensive Skilled craftsmen who build the product Significant work-in-process inventories Often warehouse work-in-process Jobs not overlapped between work centres Ad-hoc and dynamic customer orders

Operational Characteristics of the SME PE Companies

Electronics & Semiconductor Backend Fixed routings Unique machine capability at each work centre (Specialized equipment) Product layout High volume Capacity is well defined Capital intensive Highly specialized, trained operators who monitor and control process equipment Low work-in-process inventories No warehousing of work-in-process Job overlapping Relatively stable customer requirement, with forecast provided to guide long term capacity planning

Overview of the Dynamic Operations and Manpower Scheduling System

Figure 1 depicts the functional blocks of our proposed dynamic operations and manpower scheduling system. It shows the interaction of the Scheduling System with the external environment. Information can be specified through the Graphical User Interface (GUI) or automatically pulled from the existing manufacturing systems through the data integration mechanisms. To prevent duplicate data entry effort and reduce data maintenance, if the data are originated from other systems like Enterprise Resource Planning (ERP) or Manufacturing Execution System (MES), they should be updated to the scheduling system through integration program. Existing business rules such as scheduling policy and operational constraints can be modeled and developed as heuristics rules to be embedded into the scheduling engine. The basic output from the system is the detailed production schedule, which can be transformed into a set of operation (e.g. dispatch list for each machine) and management (e.g. cycle time and machine utilization) reports.

The operational characteristics of the PE companies as compared to the Electronics & Semiconductor Backend companies are highlighted in Table 1. In contrast with the typical high-volume low-mix, flow shop environment in the Electronics and Semiconductor Backend environment, the PE companies operate in a low-volume, high-mix job shop environment with variable product routing. In this environment, the resources (machine and manpower) are highly flexible and capable of performing different operations. In most cases, it is quite labour intensive and it also requires skilled technician to build the product. In addition, due to its supporting role and the fact that it supports multiple industrial sectors, it is also subjected to ad-hoc, dynamic and last-minute urgent customer requests. The proposed solution is attempted to satisfy both the Manpower and Machine resource constraints. The rescheduling engine is developed to cater for the highly flexible machine resources and the

154

T. J. Chua et al

The key features of the system are the job shop scheduling engine which is able to handle the variable routing of the work orders, the manpower scheduling engine to generate staffing plan to support production schedule, the easy-to-configure scheduling rules in-

corporating the maximize machine utilization, minimize work-in-process (WIP) scheduling heuristics rules and the dynamic drag-and-drop schedule adjustment through Interactive Gantt Chart graphical user interface.

Fig. 1. Functional modules of the Dynamic Operations and Manpower Scheduling Systems.

4

4.1

RESULTS AND DISCUSSION

The design and the functionality of the proposed system has been presented and discussed with a number of companies in the PE industry. We have received favorable feedback from the end user communities. A consortium-approach has been adopted in order to lower the entry cost for the deployment due to low affordability of the PE companies. Up to 70% of the estimated R&D effort and cost will be co-funded by the government. The consortium members will share out the remaining 30% cost among themselves. Development effort is currently on-going in parallel for both the manpower scheduling and dynamic rescheduling modules. Some initial results of the respective modules are presented in the following sections.

155

Manpower Scheduling Algorithms

The manpower scheduling algorithms are activated after the production schedule has been generated by the job shop scheduling engine. The objective is to minimize operating cost, given a set of operators having different skill sets and job grades, and having to work on different operating shift patterns. The execution sequences of the manpower scheduling algorithms are: Step 1: Obtain all available skill sets defined in the system. Step 2: Obtain the mapping relationship between product and skill sets. Step 3: Retrieve job-machine assignment based on the latest schedule generated by the job shop scheduling engine Step 4: Obtain operator list and build the time line for each operator based on their working shift pattern.


on heuristics. The future plan is to implement optimization algorithm based operation research techniques.

Step 5: Based on all above information, assign operator to machine according to the following steps: (a) Based on the schedule result (job assignment), calculate the required skills, required operator number and operation time on each working machine. (b) Based on the working calendar of each operator and the mapping relationship between operator and product, calculate available operator number for each skill set required by each machine. (c) Based on the preferred job grade for each skill set and score of each operator, reserve operator for each machine with maximum score and job grade less than or equal to preferred job grade. The basic idea behind the computation of the operator’s score for each skill set is to obtain the difference between the required skills and available skill. The larger difference, the higher score for this skill set. (d) Assign job to operator based on the reservation and consider overtime requirements if job cannot be assigned with the existing list of operators. Figure 2 depicts the product schedule highlighting both the machine and operator assignment results. The current manpower scheduling algorithm is based

4.2

Dynamic Operations Scheduling

The current effort on the dynamic operations scheduling module has been focused on the development of the interactive gantt chart with drag-and-drop capability. Figure 3 shows the design of the gantt chart. The Gantt chart is the standard format for displaying a schedule graphically. It consists of a horizontal bar chart with time as the horizontal axis and the machine resources as the vertical axis. Individual work order assignment is displayed as horizontal bar in the chart, indicating the time at which the operation begins and ends. The developed gantt chart graphical user interface allows the end user to drag the ad-hoc urgent orders from the work order list in the upper panel of the user interface and drop into the generated production schedule in the lower panel. Any constraint violation will be checked and the end user will be prompted for decision. Additional features include ‘freezing’ of the assigned work orders and partial and complete rescheduling. The heuristics and optimization algorithms for reactive scheduling are currently being designed and developed.

Fig. 2. Dispatching result showing the operator assignment.

156

T. J. Chua et al

Figure 3. Interactive gantt chart for Dynamic Schedule Adjustment.

5

CONCLUSION

The production environment in the PE industry is characterized by a large and diverse product mix with smaller batch size and shorter lead times. To compete effectively, excellent customer services must be delivered while keeping the price low. Therefore, it is important to be able to control and utilize the resources (machine and manpower) effectively, to improve delivery performance, and reduce setup and production time losses. Many of these companies are feeling the need for a computerized production scheduling tool to assist them in managing the complex set of inter-related constraints. In this paper, the initial design and development of a prototype system for operations and manpower scheduling to facilitate management of this complex and dynamic environment, have been presented and discussed. 6

INDUSTRAL SIGNIFICANT

Based on industrial survey and visits to a number of SMEs in the PE industry, including Allied Technologies, ACP Metal Finishing, II-VI Singapore, Getech Precision Tooling, Endela Trading & Mfg, Vigor Precision, Fong’s Engineering & Manufactur-

157

ing, Gee Sheng Machinery & Engineering, M.C. Packaging, Long Tech Engineering, Leo Industries, Hitech Heat Treatment Services, PLC Industries, Infiniteglo, PLC Tech, JCS Group, Holy Engineering, Wah Son Engineering, etc. It was observed that almost all companies visited are still adopting Excel-based manpower planning & scheduling. A dynamic operations and manpower scheduling system uses advanced programming techniques to improve and optimize production schedule, to allow the company to achieve pre-defined objectives, such as allowing improvements on the delivery performance without raising inventory levels, or maximize the plant throughput is crucial for PE SME companies. Through the PE-COI initiative, we have attempted to reach out to these companies through a cost sharing model of consortium approach and potential funding from the SPRING Singapore, this will help to lower the entry cost and make the solution more affordable to these SMEs.


[16] W.B. Henderson, “Heuristic methods for telephone operator shift scheduling: an experimental analysis”, Mgmt. Sci., vol. 22, pp. 1372-1380, 1976. [17] S.E Bechtold, L.W. Jacobs, “Implicit modeling of Flexible break assignments in optimal shift scheduling”, Mgmt. Sci., vol. 36 (11), pp. 1339-1351, 1990. [18] T. Aykin, “A comparative evaluation of modeling approaches to the labor shift scheduling problem”, Eur. J. Oper. Res., vol. 125, pp. 381-397, 2006. [19] B. Grabot and A. Letouzey, “Short-term manpower management in manufacturing systems: new requirements and DSS prototyping”, Comput. Ind., vol. 43, pp. 11-29, 2000. [20] M. Belbin, Management teams: why they succeed or fail, London: Heineman, 1981. [21] E. Berne, The structure and dynamic of organisation and groups, New York: Ballantine, 1984. [22] K. Bursic, “Strategies and benefits of the successful use of teams in manufacturing”, IEEE Trans. Eng. Manage., vol. 39(3), pp. 1606-1615, 1992. [23] H.K. Alfares, J.F. Bailey, “Integrated project task and manpower scheduling”, IIE Trans., vol. 29(9), pp. 711 -717, 1997. [24] R.N. Burns, M.W. Carter, “Workforce scheduling with cyclic demands and days-off constraints”, Manage Sci., vol. 24, pp. 161-167, 1985. [25] H. Emmons, “Workforce scheduling with cyclic requirements and constraints on days off, weekends off and workstrech” IIE Trans., vol. 17, pp. 8-16, 1987. [26] R. Hung, “Multiple-shift workforce scheduling under the 3-4 workweek with different weekday and weekend labor requirements”, Manage Sci., vol. 40, pp. 280-284, 1994. [27] R.S. Russell, P.Y. Huang, and Y. Leu, “A study of labor allocation strategies in cellular manufacturing”, Decision Sciences, vol. 22(3), pp. 594-611, 1991. [28] G.T. Wirth, F. Mahmoodi, C. T. Mosier, “An investigation of scheduling policies in a dual-constrained manufacturing cell”, Decision Sciences, vol. 24(4), pp. 761-788, 1993. [29] C. Yano, and R. Rachamadugu, “Sequencing to minimize work overload in assembly lines with product options”, Mgmt Sci., vol. 37(5), pp. 1-15, 1991. [30] A.G., Lagogimos and V., Leopoulos, “Greedy heuristic algorithms for manpower shift planning”, Int. J. Prod. Econ., vol. 68(12), pp. 95-106, 2000.

REFERENCES [1] J. Sun, D. Xue, “A dynamic reactive scheduling mechanism for responding to changes of production orders and manufacturing resources”, Comput. Ind., vol. 46, pp. 189-207, 2001. [2] D. Jürgen, “Iterative Improvement Methods for Knowledge-based Scheduling”, AI Communications, vol. 8(1), pp. 20-34, 1995. [3] J. Dorn, M. Girsch, G. Skele, “Comparison of Iterative Improvement Techniques for Schedule Optimization”, Eur. J. Oper. Res., vol. 94(2), pp. 349-361, 1996. [4] A.S. Raheja and V. Subramaniam, “Reactive Recovery of Job Shop Schedules – A Review”, Int. J. Adv. Manuf. Technol., vol. 19, pp. 756-763, 2002. [5] G.E. Vieira, J.W. Herrmann, E. Lin, “Rescheduling manufacturing systems: a framework of strategies, policies, and methods”, J. Scheduling, vol. 6(1), pp. 39-62, 2003. [6] Y.P. Ran, N. Roos, J. Herik, “Methods for Repair Based Scheduling”, Proc. 21st Workshop of the UK Planning, 2002. [7] A. Gallagher, T.L. Zimmerman, S.F. Smith, “Incremental Scheduling to Maximize Quality in a Dynamic Environment”, Am. Assoc. Artificial Intelligence, pp. 222-231, 2006. [8] P. Wirojanagud, J. Fowler, and E. Gel, “Workforce Planning in Semiconductor Manufacturing”, 3rd Int. Conf. Modeling Analysis of Semicond. Manufact., October 2005, pp. 6-7. [9] B. Grabot, and A. Letouzey, “Short-term manpower management in manufacturing systems: new requirements and DSS prototyping”, Comput. Ind., vol. 43, pp. 11-29, 2000. [10] H.C. Lau, “On the Complexity of Manpower Shift Scheduling”, Computers Ops. Res., vol. 23(1), pp. 93-102, 1996. [11] M.K. Wong, and S.G. Ng, “A Manpower Recovery Decision Support System with Time Window and Skill Restrictions”, Int. Conf. Service Sys. Service Mgmt., Chengdu, China, 9-11 June 2007, pp. 1-6. [12] L.F. McGinnis, W.D. Cilver, and R.H. Deane, “Oneand Two-phase Heuristics for Workplace Scheduling,” Comput. Ind. Eng., vol. 2, pp. 7-15. 1978. [13] L.C. Edie, “Traffic delays at toll booths”, Ops. Res., vol. 2, pp. 107-138, 1954. [14] G.B. Dantzig, “A Comment on Edie's traffic delay at toll booths”, Ops. Res., vol. 3, pp. 339-341, 1954. [15] W.B. Henderson, and W.L. Berry, “Determining Optimal Shift Schedules For Telephone Traffic Exchange Operators”, Paper No. 507, Krannert Graduate School of Industrial Administration, Purdue University, West Lafayette, Indiana, 1975.

158