A System of Systems Framework - IEEE Xplore

256

IEEE SYSTEMS JOURNAL, VOL. 5, NO. 2, JUNE 2011

Exploring the Dimensions of Systems of Innovation Analysis: A System of Systems Framework Ali Mostafavi, Student Member, IEEE, Dulcy M. Abraham, Daniel DeLaurentis, Member, IEEE, and Joseph Sinfield

Abstract—The objective of this paper is to propose an analysis framework based on the System of Systems approach to overcome existing methodological problems in System of Innovation studies. The concept of System of Innovation has been an important focus of innovation policy studies over the last decade. These studies have concentrated on structuring theoretical frameworks to assess the determinants of innovation processes using systems analysis principles. Despite about 20 years of research, System of Innovation has yet to become a theoretical framework. This paper reviews the relevant literature in an attempt to identify the challenges that System of Innovation studies face in structuring theoretical frameworks. The primary reason for the challenges appears to be that Systems of Innovation have been analyzed as monolithic systems when, in reality, Systems of Innovation are Systems of Systems, which have different features than monolithic systems. Different dimensions of assessment of System of Systems analysis are therefore required in System of Innovation studies. Three dimensions of System of Innovation analysis (definition, abstraction, and modeling) are introduced here to provide an analysis framework for Systems of Innovation studies. The proposed system-of-systems-based analysis framework (called Innovation System of Systems) would resolve the methodological challenges that System of Innovation studies are confronting in developing theoretical frameworks. Thus, it is capable of being tested by other researchers in the area of Systems of Innovation to advance the state of knowledge. Index Terms—Infrastructure, innovation policy analysis, intelligent transportation systems, system of innovation, system of systems.

I. INTRODUCTION HE objective of this paper is to propose an analysis framework based on the System of Systems approach to overcome existing methodological problems in System of Innovation (SoI) studies in order to be possibly tested by other researchers and thus advance the state of the knowledge. The concept of SoI has been an important focus of innovation policy studies over the last decade. These studies have concentrated on structuring a theoretical framework to assess the determinants of innovation processes using systems analysis principles. The SoI concept is defined as “all important economic, social, political, organizational, institutional, and other factors that influence the development, diffusion, and use of innovation” [1]. The constituents of SoIs include organizations and institutions and the relationships among the components. Organizations are players or actors, and institutions are the sets of common habits, norms, routines, established practices, rules, or laws that regulate the

T

Manuscript received July 29, 2010; revised December 12, 2010; accepted February 17, 2011. Date of publication April 21, 2011; date of current version May 25, 2011. A. Mostafavi, D. M. Abraham, and J. Sinfield are with the School of Civil Engineering, Purdue University, West Lafayette, IN 47907-2051 USA (e-mail: [email protected]; [email protected]; [email protected]). D. DeLaurentis is with School of Aeronautics and Astronautics, Purdue University, West Lafayette, IN 47907-2045 USA (e-mail: [email protected]). Digital Object Identifier 10.1109/JSYST.2011.2131050

relationships and interactions between the players and actors. Activities in SoIs are the functions of the players that influence the development, diffusion, and use of innovation [1]. Research, development, and implementation are examples of activities in SoIs. Three SoIs have been identified in the literature: National SoIs (NSoI) ([2]–[4]) Regional SoIs (RSoI) ([5]–[7]), and Sectoral SoIs (SSoI) [8], [9]. Each SoI level focuses on different factors, players, and interactions, all of which will be discussed in more detail later in this paper. Thus far, SoI has failed to be recognized as a “theoretical” framework designed to explain the behavior of a system, despite decades-long efforts to analyze it as such. A theoretical framework is a structure that can hold or support a theory constructed to explain phenomena and recognize patterns (e.g., relationships, events, or behaviors) in a setting. The reason that SoI has remained an “under-theorized” approach has been partly attributed to insufficient case study research to compare different SoIs [10]. Another factor that has hindered progress can be attributed to the methodological issues that exist in studies of SoI [11], making it difficult to search for propositions. Methodology is the branch of logic that deals with the general principles of the formation of a discipline. Addressing methodological issues would clarify the aspects of SoIs (such as contexts, categories, and levels of abstraction) on which case study research should focus. Coupling case study research with a well-structured analysis framework that captures the dimensions of SoI analysis to address methodological issues would finally lead to theoretical frameworks for innovations. A Sue analysis framework should be consistent with the traits of SoI and should be capable of capturing its analysis dimensions, such as definition and abstraction. The SoI approach was first developed by Freeman [2] to adopt systems analysis for understanding innovation processes. SoI was thought to consist of components and the relationships between components as a monolithic system. However, SoI is a System of Systems (SoS), which is defined as a set of individual monolithic systems that interact to obtain functionalities that cannot reside in the individual systems. A SoS encompasses different traits compared to a monolithic system, whose traits include an operational dependence between the components, a hierarchical structure, and static boundaries. There is dependence between the components during the interactions to fulfill the system’s function while the individual subsystems in a SoS are operationally and managerially independent. The outcome of the individual subsystem interrelationships in a SoS is an emergent behavior that cannot reside in any subsystem, but rather “emerges” due to the interactions between/among the subsystems. Analysis of SoSs therefore consists of different dimensions compared to monolithic systems, and treating SoIs as monolithic systems will result in a failure to capture the required traits and dimensions of analysis.

1932-8184/$26.00 © 2011 IEEE

MOSTAFAVI et al.: EXPLORING THE DIMENSIONS OF SYSTEMS OF INNOVATION ANALYSIS

In the following sections of the paper, different SoIs first are discussed and the various characteristics of SoIs are identified to assess the relevance of the SoS approach. Then, the architecture, the lexicons, and the principles of SoS are introduced to identify the state-of-the-art advancements in the concept. Finally, an analysis framework for SoI analyses is provided. II. METHODOLOGICAL CHALLENGES SYSTEM OF INNOVATION

OF

The study of innovation has been a central point of attention in empirical economics literature for more than 30 years. The innovation studies assess different aspects of innovation and can be divided into two broad categories. The first group of innovation studies assesses innovation processes at a micro level. These studies focus on structural analysis of innovation processes to create propositions regarding invention and diffusion of different types of innovations (e.g., architectural, generational, incremental, radical, and disruptive innovations) to provide prescriptions for innovators. Despite their invaluable effectiveness for evaluating innovations structurally, this group of innovation studies (e.g., Christensen [12] and Henderson and Clark [13]) does not provide propositions for innovation policy-making. Innovation policy-making assesses the organizations, interactions, and factors affecting the development and diffusion of innovation and making policies to enhance innovation processes. Innovation policy-making requires a macro-level analysis of the organizations, interactions, and factors affecting the development and diffusion of innovation. The second group of innovation studies considers innovation processes at a macro level. These studies evaluate the roles and activities of different organizations and the interactions among the organizations and their effect on innovation processes. This group of innovation studies introduced a SoI approach to provide propositions for innovation policy-making. The focus of this paper is on SoI studies. This group of innovation studies, their major contributions, and the existing challenges they encounter are discussed in the remainder of this section. The notion of SoI has been an important focal point of innovation policy studies over the last two decades. These studies have focused on structuring theoretical frameworks to evaluate the determinants of innovation processes using the principles of systems analysis. Freeman defines National SoIs (NSoI) as “the network of institutions in the public and private sectors whose activities and interactions imitate, import, modify, and diffuse new technologies” [2]. In the NSoI definition Freeman tacitly emphasized the fact that the activities of organizations “contribute” to innovation development and diffusion [2]. After assessing NSoIs in Japan, Germany, and East Asian and Latin American countries, Freeman concluded that different countries have different NSoIs to emphasize the different institutional arrangements in nations [14]. Approaches for NSoI studies have focused on the “learning and knowledge flow” in the system [3], emphasizing the significance of “communication” between the system constituents. The players of NSoIs include but are not limited to governments, corporate research and development (R&D) divisions, universities, and venture capital organizations. Players and institutions of NSoIs are not fixed over time due to the “evolutionary process” involved in NSoIs [4]. The evolutionary nature of SoIs is due to changes of players, activities, and interactions over time. New players might join

257

or existing players might leave a SoI. Activities of and interactions between players might change due to internal factors such as the advent of a new player or external factors such as global economic conditions. Sectoral SoIs (SSoIs) are defined as “networks of agents interacting in a specific technology area under a particular institutional infrastructure for the purpose of creating, diffusing, and utilizing technology focused on knowledge, information, and competence flow.” [9] Breschi and Malerba delineate SSoIs as “the specific clusters of the firms, technologies, and industries involved in the generation and diffusion of new technologies and in the knowledge flow that takes place among them.” [8] SoIs in the telecommunication sector or bio-energy sector are examples of SSoIs. The definition captures most aspects of SSoIs, implying the “independence” of the agents while they “contribute” to the whole system purpose of innovation development and diffusion. Players of SSoIs include but are not limited to consumers, entrepreneurs, scientists, technical associations, government agencies, and universities. SSoIs perform independent functions in different sectors. Finally, Regional SoIs (RSoI) have come into play as a response to “the perceived importance of the local supply of managerial and technical skills, accumulated tacit knowledge, and knowledge spillover” [11]. Similar to NSoI and SSoI, the RSoI definitions implicitly place emphasis on the communication and independence of the players. This can be clearly understood in the definition offered by Group de Recherche European Sur les Milieux Innovateurs [15] of subnational (local) SoIs, which emphasizes informal relationships (communication) and a collective learning process (evolving nature). Each of the examined SoIs has its system boundaries: NSoI is nation-bounded, RSoI is region-bounded, and SSoI is sector-bounded but not necessarily nation bounded [11]. SoI has yet to become a theoretical framework partly due to some methodological challenges that have not been addressed thus far. Generally, there are five methodological challenges impacting SoI analysis: the lack of a consistent methodological lexicon, the need to address the impact of the interrelationships between different SoIs, the need to identify the appropriate level of analysis, the need for making ex-ante predictions, and the dynamics of systemic boundaries [11]. A comprehensive analytical framework for SoI analysis should not only be consistent with the characteristics of SoIs, such as the operational and managerial independence of the players, but should also address the existing methodological challenges [11], in order to pave the way for case study research projects to develop innovation theories [16]. The primary reason for the above-mentioned challenges appears to be that SoIs have been analyzed as monolithic systems when, in reality, SoIs are Systems of Systems (SoSs). SoSs have different features than monolithic systems. In the following section, the SoS approach is introduced as an appropriate analytical framework for designing SoI studies. Then, the traits of SoSs in SoIs are evaluated and the dimensions and elements of SoI analysis are proposed. III. SYSTEM OF SYSTEMS The first step in systems analysis is to determine the type of the system (e.g., monolithic, complex adaptive systems or system-of-systems). Different types of systems have different

258


SYSTEMS

AND

TABLE I SOS ANALYSIS TRAITS [20]

traits. SoS traits will be presented in this section in order to determine whether SoI can be characterized as SoS. Rechtin defines a monolithic system as “a set of different elements so connected or related so as to perform a unique function not performable by the elements alone” [17]. Along the same lines, DeLaurentis and Callaway [18] characterize a SoS as “the combination of a set of different systems [that] forms a larger system of systems that performs a function not performable by a single system alone.” Maier defines a SoS as “an assemblage of components which individually may be regarded as systems, and which possesses two additional properties: operational independence of components and managerial independence of the components” [19]. SoSs have different traits compared to monolithic systems. Gorod et al. [20] summarizes the major traits of monolithic systems and SoS analysis as shown in Table I based on the works of Keating et al. [21], Maier [19], Delaurentis [22], and Bar-Bar-Yam et al. [23]. As shown in Table I, analysis of SoSs is different from that of monolithic systems. Therefore, failing to recognize the system type could lead to analytical and methodological problems. Since the emergence of SoS, different categories of its distinguishing traits have been introduced. Maier [24] cites geographical distribution, emergent behavior, evolutionary development, operational independence, and managerial independence as distinguishing characteristics. Boardman and Sauser [25] cite autonomy, belonging, connectivity, diversity, and emergence as distinguishing characteristics that are generally the same as those introduced by Maier [24]. Maier also concluded that managerial and operational independence are primary (necessary) conditions for a SoS setting, and emergence, evolutionary behavior, diversity (geographical distribution), and connectivity are secondary conditions. If a setting lacks the primary conditions, he contends, it cannot be considered a SoS even if it has the secondary characteristics. Maier [19] defines operational independence as follows: “if the SoS is disassembled into its component systems, the components systems must be able to usefully operate independently” and managerial independence as follows: “the component systems not only can operate independently, they do operate independently.” A SoS performs functions and carries out purposes that do not reside in any component system. These behaviors are the emergent properties of the entire SoS and cannot be localized to any component system. The principal purposes of the SoS are fulfilled by these behaviors [24]. SoSs can be characterized from another aspect as: directed, collaborative, and virtual [24]. Identifying whether a SoS is directed, collaborative, or virtual enables a better understanding of SoS dynamics and structure. A directed SoS has a centrally managed system to assure that

the SoS function is performed, whereby the component systems functions are subordinated to the SoS function. A collaborative SoS lacks a centrally managed system. It is the collaboration between the key component systems and the players which provides an enforcement mechanism to fulfill the SoS function. Finally, in a virtual SoS, there is not a centrally agreed-upon purpose for the SoS. In the following section, these SoS traits are reflected in the SoIs in order to address the challenges of current SoI studies using a SoS analysis framework. IV. ADDRESSING SOI CHALLENGES USING THE SOS APPROACH SoS taxonomies and architecture principles can be applied to identify the dimensions of a SoI analysis framework since SoIs have been determined to be SoSs as described in Section III. The first step is to assess whether a SoI can be considered as a SoS. Maier’s criteria form the basis of the assessment, namely, the players in every SoI are individual systems that are operationally independent. For instance, public and private institutions (i.e., companies, governmental organizations, etc.) that are the players of NSoIs can operate independently. This is also the case in SSoIs and RSoIs. Not only can the players operate independently, but they also operate autonomously, which verifies the operational and managerial independence of the component systems in a SoI. In addition to operational and managerial independence, the players of every SoI are geographically distributed. Even in NSoIs, players operate in different geographic locations. Geographical distribution of the players can also be recognized in SSoIs and RSoIs. One of the indicators of geographical distribution is the existence of communication between the component systems. Lundvall [3] cites that linkages between the players of a SoI can be found at the interfaces in the form of learning and knowledge flow (communication). Consequently, it can be concluded that SoIs are geographically distributed. As implied in the definition of SoI, the emergent function of a SoI is to “imitate, import, modify, and diffuse new technologies (innovation)” [2]. This emergent function cannot reside in any component systems of SoIs. Moreover, SoIs have been developed over time. The interactions between the players, as well as the players of every SoI, evolve with time. “They (SoIs) employ historical and evolutionary perspectives, rendering the notion of optimality irrelevant” [10]. Thus, SoIs have all of the five distinguishing SoS characteristics and can be regarded as such. Recent studies clearly reflect the SoS nature of SoI: Edquist [1] defines the constituents of SoI as organizations and institutions and the relations among the components. Organizations are players or actors, and institutions are the sets of common habits, norms, routines, established practices, rules, or laws that


SO I

259

TABLE II SOS STRENGTHS

CHALLENGES AND

regulate the relations and interactions between the players and the actors. In fact, they are the “rules of the game.” Activities in SoIs are functions of the players that influence the development, diffusion, and use of innovation. Institutions are the component systems’ operational function. Finally, the activities constitute the players’ roles for fulfilling the emergent function of the SoI. To emphasize the advantages of using SoS lexicons and principles for enhancing the SoI framework, the SoI challenges cited by Chang and Chen [11] are listed against the corresponding robustness of SoS analysis in Table II. An examination of Table II reveals that the SoS approach is appropriate for designing SoI studies, thereby providing the foundation for a theoretical framework. In the following section, the dimensions of a SoI analysis are proposed in a new framework called Innovation System of Systems (I-SoS). A case study is used to illustrate the I-SoS framework dimensions. V. I-SOS ANALYSIS FRAMEWORK To illustrate the application of the I-SoS framework, this section uses an example of innovation policy-making in transportation infrastructure in the area of intelligent transportation systems (ITS). ITS enhance transportation energy consumption, mobility, and safety through the integration of innovative communications technologies (e.g., deployment of wireless communication, computational technologies, sensing technologies, and inductive loop detection) into the transportation infrastructure. There are six categories of ITS, which include: Advanced Traveler Information Services (ATIS), Advanced Traffic Management System (ATMS), Advanced Vehicle Control Systems (AVCS), Commercial Vehicle Operations (CVO), and Advanced Public Transport (APT) [26]. The components of ITS include digital broadcast, digital cellular data networks, wide area digital networks, short range vehicle beacons, and navigation [27]. Development and diffusion of innovative ITS in the nation’s transportation infrastructure is vital since “the nation’s outdated transportation systems negatively impact the international competitiveness of the U.S. economy and the quality of life of the citizens” [30] and requires policies (e.g., obligatory and regulatory policies) to enhance development and diffusion of innovation. An example of such policies is the ITS Strategic Research Plan, which was released in 2010. The subject of the plan is transforming the nation’s transportation system through connectivity using ITS. The policy intends to facilitate development and diffusion of innovative ITS. Innovation policy-making without a structured analysis of different dimensions of the problem might not yield the desired outcomes. The question to be answered is: “What dimensions

should be assessed in the analysis of innovation in ITS in transportation infrastructures for policy-making purposes?” Prior studies (e.g., [27]–[29]) have not considered ITS in such a context. While the purpose of these studies is to explore the physical architecture and components of ITS using SoS, the objective of the case study of ITS in this paper is to provide an analytical framework to identify the dimensions of innovation analysis. The ITS innovation policy analysis is a SoI problem since the process of development and diffusion of ITS innovation cannot be performed by individual organization(s) and is the emergent result of activities and interactions among different organizations. Thus, it can be assessed using the I-SoS framework. If the existing SoI approach is used, ITS would be considered within either the communication technology sectoral SoI or the transportation infrastructure sectoral SoI at either the national or the subnational level. Such an approach may fail to: i) address the impact of the interrelationships between different SoIs (e.g., failing to assess the players of the transportation infrastructure sectors at different levels who affect the innovation process but are not included in the communication technology sectoral SoI); ii) identify the appropriate level of analysis (e.g., focusing on national level players and interactions while failing to consider subnational players); and iii) consider the dynamics of systemic boundaries (e.g., failing to assess the dynamics among communication technology sectoral SoI and national SoI boundaries). However, using the I-SoS framework, the interrelations between the different SoIs are taken into account, whereby the players of national/subnational and communication technology/ infrastructure sectoral SoIs can be included in the analysis. Appropriate levels of analysis are identified by assessing the sub national, national, and global players and the interactions within and across sectoral SoIs. The dynamics of systemic boundaries also are considered (e.g., by considering the interrelations between the players in the communication technology sectoral SoI and the infrastructure sectoral SoI). Fig. 1 summarize the elements of analysis considered by the I-SoS framework versus those considered using the traditional SoI approach. The dimensions of analysis for ITS innovation are identified in the remainder of this section using the I-SoS framework. The ITS example in this section is intended to illustrate the application of the I-SoS framework to identify the dimensions of analysis, rather than analyzing the ITS innovation itself, which is beyond the scope of this paper. Fig. 2 summarizes the I-SoS framework dimensions. As shown in Fig. 2, each dimension consists of different analysis elements.

260

Fig. 1. SoI analysis versus I-SoS analysis.

The three dimensions (definition, abstraction, and implementation) of the I-SoS framework are defined as follows. 1) Definition: The definition dimension consists of defining the context, categories, levels, and barriers elements: Context defines the boundaries of the SoI analysis. SoI studies range from analysis of technology changes to knowledge development and diffusion and learning. For ITS analysis, the context is assessment of development and diffusion of innovative ITS within the U.S. transportation infrastructure. There are different innovation categories, such as product, process, and technology innovations. In the context of ITS, the innovation category would be technology innovations (e.g., innovative communication and sensing technologies in ITS, such as wireless communication, computational technologies, sensing technologies, and inductive loop detection). Appropriate levels of analysis (i.e., global, regional, national, subnational, or organizational) are defined through level elements. As discussed earlier, SoIs have been studied at three levels in the literature: national, regional, and sectoral. In addition, some researchers (refer to [31] and [32]) claim that supernational SoIs are emerging. Sectoral SoIs could also have national and global domains. Besides, at the lowest level, organizations are also SoIs. Players at the organizational level include employees and managers. This level of analysis considers activities and institutions among the players in an organization, and innovation in the organization is the emergent result of these activities and institutions. Thus, the I-SoS framework , Local or Sub-national considers five levels: Organizational , National Regional , and Global . Sectors may exist within each of these levels. Fig. 3 shows the levels in the I-SoS framework. Each level of analysis consists of specific players and institutions within each sector. Sectors of interest in the ITS case include transportation infrastructure and communication technology. Depending on the intention of a study, the appropriate level for analysis would be different. As a case in point, global technological changes in the context of ITS may be of interest for a study. In this case, the appropriate level of analysis would be the global level. In another case, the diffusion of ITS technological changes in a country’s infrastructure may be of interest, and hence, the national or subnational levels would be appropriate levels of analysis as shown in Fig. 4. Whatever


the level of analysis might be, the effects of the within- and cross-level players and their interrelationships should be addressed in the assessment. For instance, in the ITS case, the interactions between the players at the national level, such as the Federal Highway Administration (FHWA) and the Research and Innovative Technology Administration (RITA) within the U.S. Department of Transportation (DOT), and the interrelations between players across levels (e.g., interrelations between researchers at the subnational level and USDOT at the national level) and across sectors (e.g., interrelations between firms in the communication technology sector and RITA in the transportation infrastructure sector). In Fig. 4, the players and interdependencies among the players are shown within and across subnational (beta-level) and national (gamma-level) levels and within and across the communication technology and transportation infrastructure sectors for ITS innovations in the U.S. A barrier element takes obstacles in the analysis into account. Obstacles vary in different studies. An obstacle that one would expect in an analysis of ITS innovation might be the large number of players and the heterogeneity of their activities and the institutions that make the analysis very complex. For instance, in the ITS case, players include but are not limited to contractors, engineering firms, communication and sensing technology and equipment suppliers, and public agencies such as FHWA and RITA within the USDOT) which perform different activities (e.g., policy-making and investment by USDOT, research and development by communication and sensing technology firms, and implementation by contractors). It is a complex problem to take all of the players and their interdependencies into consideration in an analysis. 2) Abstraction: In the abstraction phase of the SoI analysis framework, the players, institutions, activities, networks, and resources are identified. Players, in the case of ITS, in the transportation infrastructure sector include but are not limited to USDOT, FHWA, state DOTs, and engineering and construction companies. In the communication technology sector, they include but are not limited to communication technology and equipment suppliers (e.g., IBM and Siemens), associations (e.g., Intelligent Transportation Association of America), and researchers. Fig. 4 illustrates the players discussed. Each of these players represents a different institution, and acts differently. Institutions can be classified based on how they are governed, either through formal laws or through organizational heuristics (formal versus informal institutions). An example of institutions in the former classification in the case of ITS include the Research and Special Programs Improvement Act passed by the U.S. Congress in 2005 which created the Research and Innovative Technology Administration (RITA) within USDOT. RITA coordinates the department’s multimodal research and education programs and advances the deployment of cross-modal technologies into the transportation system [33]. An example of informal institutions would include the lessons learned and best practices. Drivers and inhibitors are activities in SoIs that influence the development, diffusion, and use of an innovation. In the case of ITS, the activities include but are not limited to research (e.g., basic, developmental, engineering, etc.), investment, and implementation (construction). The activities of the players vary within and across levels. For instance, the primary activity of the federal agencies (FHWA and RITA


261

Fig. 2. Dimensions of analysis in I-SoS.

Fig. 3. Levels of analysis in I-SoS.

within the USDOT) as a national level player (i.e., gamma level as shown in Fig. 4) would be developing appropriate policies, making investment decisions, and implementing research. On the other hand, the activities of communication and sensing technology firms in the communication technology sector and at the subnational level (beta level as shown in Fig. 4), include ITS technological innovations research. RITA coordinates the research programs of USDOT and is charged with advancing the deployment of innovative technologies to improve the U.S. transportation system. Furthermore, the Enterprise Program is another example of an activity to facilitate ITS research activities. The Enterprise Program is a pooled-fund study with member agencies from North America and Europe. Its main purpose is to use the pooled resources of its members, private sector partners, and the U.S. federal government to develop, evaluate, and deploy ITS. The Transportation Technology Innovation and Demonstration (TTID) program is another example of activities by USDOT to facilitate ITS deployment. The activities of local engineering companies (as subnational level players, depicted at the beta level in Fig. 4) include but are not limited to research and development. Interdependencies

between the players of SoIs can be formal or informal. Formal interdependencies include transfer of physical resources between the players, while informal interdependencies include communication and knowledge transfer among the players. In the ITS case, for instance, formal interdependencies between USDOT and researchers include transfer of financial resources, while informal interdependencies between communication technology firms (inter-firm networks) include knowledge transfer and communication as shown is Fig. 4. Formal interrelations and informal interrelations are shown with solid lines and dotted lines, respectively, in Fig. 4. Another element of the abstraction dimension of the analysis framework pertains to the SoI’s networks. Different networks have been identified in the SoIs literature: technological networks [31], inter-firm networks [33]–[35], social networks [11], learning networks [36], supplier-user networks, pioneers-adopters networks [37], problem solving networks, and informal networks [38]. Different networks exist within and across different levels. In the case of ITS, inter-firm networks (such as the network of communication technology firms) and financial resources networks are examples of networks in the transportation infrastructure and communication technology sectors within the subnational (beta) and national (gamma) levels, as shown in Fig. 4. Either set of networks as a whole or components and links in networks can be the subjects of ITS innovation policy analyses. In addition, different innovation network paradigms can be the subject of the analyses. For example, networks (i.e., inter-firm networks) within an ITS System of Innovation can be studied within the paradigm of open versus closed innovation [39] to identify the level of control firms should exercise over new ideas to facilitate emergence of innovation in ITS. Resources are an important element of the abstraction dimension. Human resources ([40], [42]) and capital and competence sources [42] are the major identified resources. The types of sources exploited by various players at different levels are different, depending on the type of innovation. In the ITS case, for instance, federal agencies such as USDOT (at the gamma level) rely on their financial resources, while communication and sensing technology firms (at the beta level) rely on their

262


Fig. 4. Players and interdependencies in analysis of ITS innovation policy.

research and development and human sources to perform their innovation activities. 3) Implementation: Modeling objects, methods, classifications, and data are the elements of the implementation dimension of the archetype that facilitate predictions for innovation policy purposes. The object represents what is to be included in the model (i.e., the players, networks, institutions, interdependencies, etc.). The methods define the modeling tools and techniques. The data include the relevant pieces of information required for building the model. Classification is the process of analyzing the model outcomes, exploring the propositions, and interpreting the propositions. Unlike the other two dimensions, the implementation dimension has not been addressed in SoI studies to date. The objective of implementation is to understand the potential emergent outcomes of a policy in a SoI to be used for innovation policy-making purposes. The implementation dimension is not about predicting the emergent outcome of an innovation policy but is about assessing probabilities

and possibilities. The implementation dimension integrates the three dimensions of the SoI analysis archetype and facilitates developing a dynamic theoretical framework of SoI analysis to answer the SoI major questions (e.g., “Which activities are important for the development and the diffusion of specific innovations? Which institutional rules influence the organizations in carrying out these activities?”) [10]. In the case of ITS, based on the level of abstraction, different modeling techniques can be used. For instance, if a study’s level of abstraction is at the global level (e.g., assessment of global infrastructure technological changes), the Systems Dynamics (SD) method would be an appropriate modeling technique since the SD method deals with aggregates located at the highest level of abstraction. If the level of abstraction is at subnational level (e.g., assessment of the interdependencies between construction firms to study diffusion of a new technology), however, agent-based modeling (ABM) would be an appropriate modeling tool since ABM deals with modeling independent agents


263

TABLE III ILLUSTRATION OF THE I-SOS FRAMEWORK IN CASE OF ITS INNOVATION POLICY

Research and Innovative Technology Administration

and their interdependencies. The models can be used to generate information for policy-makers, such as the emergent outcome of player interactions as a result of an innovation policy. In fact, failing to take all the analysis dimensions into account might lead to ineffective policies. Table III summarizes the ITS illustration, and Fig. 4 shows the players and interdependencies in the case of ITS. Not all players and interdependencies are shown in Fig. 4. Different shapes represent different players within and across different levels and sectors, and the different links represent the different types of interdependencies (e.g., formal versus informal) between players. Identification and assessment of the types of players and their activities, the types of interdependencies between players, different networks in the SoI, etc. can be the subject of various future research studies on ITS innovation policy analysis. The illustrative case of ITS highlighted the dimensions and elements of analysis for the assessment of ITS innovations in transportation infrastructure using the I-SoS framework. Unlike the previously adopted SoI approach, which assesses only the players and their activities within national, regional, or sectoral levels, the I-SoS framework considers not only the players and their activities but also the interdependencies and the effects of networks within and across different levels, thus addressing methodological challenges in the current SoI studies. VI. CONCLUSION Based on a review of the System of Innovation literature to identify the challenges of establishing it as a theoretical framework, the dimensions and elements of a new analysis framework (called Innovation System of Systems) are proposed in this paper using the System of Systems lexicon and principles. This paper contends that these challenges have been due, in part, to the treatment of Systems of Innovation as monolithic systems while they are in reality System of Systems. The components of the Innovation System of Systems analysis framework address

the challenges identified and evaluated in this paper. Three dimensions of the analysis are proposed: definition, abstraction, and implementation. These dimensions and their corresponding elements are illustrated using the case of Intelligent Transportation Systems policy-making to identify the areas that need to be analyzed by policy-makers to assess innovations in transportation infrastructures in the U.S. The Innovation System of Systems framework addresses the challenges that System of Innovation studies have faced in seeking a theoretical framework and could potentially be applied by other researchers in the area of System of Innovation studies and provides an analysis framework for development of innovation policy theories in different contexts and categories of innovations. Future studies on Systems of Innovation can apply the analysis framework and use case study research to identify important activities, institutions, interactions, and networks in a System of Innovation of interest. Such case study research can be approached by implementing within-case and cross-case comparative analyses and using other case study research steps proposed by different researchers in the literature. Examples of relevant questions for further case study investigation could include the following: 1) which activities from which organizations are important for the development and diffusion of specific innovations? and 2) which institutional rules influence the organizations in implementing these activities? Answering such questions could pave the way for robust innovation policy theories. REFERENCES [1] C. Edquist, “The SoIs approach and innovation policy: An account of the state of the art,” in National SoIs, Institutions and Public Policies Conf., Aalborg, Denmark, 2001. [2] C. Freeman, Technology Policy and Economic Performance: Lesson From Japan. London, U.K.: Frances Pinter, 1987. [3] B. Lundvall, National SoIs: Towards a Theorem of Innovation and Interactive Learning. London, U.K.: Frances Pinter, 1992.

264


[4] National Innovation Systems: A Comparative Analysis, R. R. Nelson, Ed. New York: Oxford Univ. Press, 1993. [5] P. Cooke, M. Uranga, and G. Etexbarria, “Regional innovation systems: Institutional and organizational dimension,” Res. Policy, vol. 26, pp. 475–491, 1997. [6] Evolutionary Economics and the New International Political Economy, J. De la Mothe and G. Paquet, Eds. London, U.K.: Frances Pinter, 1996. [7] H. J. Braczyk and M. Heidenreich, “Regional governance structure in a globalized world,” in Regional Innovation Systems, H. Braczyk, Ed. et al. London, U.K.: UCL Press, 1998, pp. 414–440. [8] S. Breschi and F. Malerba, “Sectoral innovation systems: Technological regimes, schumpeterian dynamics, and spatial boundaries,” in SoIs: Technologies, Organizations, and Institutions, C. Edquist, Ed. London, U.K.: Frances Pinter, 1997, pp. 130–156. [9] B. Carlsson and R. Stankiewicz, “On the nature, function and composition of technological systems,” J Evol. Econ., vol. 1, pp. 93–118, 1991. [10] C. Edquist, “Reflections on the system of innovation approach,” Science Public Policy, vol. 31, no. 6, pp. 485–489, 2004. [11] Y. C. Chang and M. H. Chen, “Comparing approaches to SoIs: The knowledge perspective,” Technol. Soc., vol. 26, pp. 17–37, 2004. [12] C. M. Christensen and M. E. Raynor, The Innovator’s Solution: Creating and Sustaining Successful Growth. Cambridge, MA: Harvard Business Press, 2003. [13] R. M. Henderson and K. B. Clark, “Architectural innovation: The reconfiguration of existing,” Admin. Sci. Quart., vol. 35, no. 1, pp. 9–30, March 1990. [14] C. Freeman, “The national system of innovation in historical perspective,” Cambridge J. Econ., vol. 19, no. 1, pp. 5–24, 1995. [15] R. Camagni, Innovation Networks: Spatial Perspectives. London, U.K.: Belhaven, 1991. [16] C. Edquist, SoIs: Technologies, Institutions and Organizations. London, U.K.: Pinter, 1997. [17] E. Rechtin, System Architecting: Creating and Building Complex Systems. Englewood Cliffs, NJ: Prentice-Hall, 1991. [18] D. DeLaurentis and R. K. C. A. Callaway, “System-of-systems perspective for public policy decisions,” Rev. Policy Res., vol. 21, no. 6, pp. 829–837, 2004. [19] M. Maier, “Architecting principles for system-of-systems,” Syst. Eng., vol. 1, no. 4, pp. 267–284, 1998. [20] A. Gorod, B. Sauser, and J. Boardman, “System-of-systems engineering management: A review of modern history and a path forward,” IEEE Syst. J., vol. 2, no. 4, pp. 484–499, 2008. [21] C. Keating, R. Rogers, R. Unal, D. Dryer, A. Sousa-Poza, R. Safford, W. Peterson, and G. Rabaldi, “Systems of systems engineering,” Eng. Manag. J., vol. 15, no. 3, pp. 36–45, 2003. [22] D. DeLaurentis, “Understanding transportation as a system of systems design problem,” presented at the 43rd AIAA Aerosp. Sci. Meet., Reno, NV, 2005. [23] Y. Bar-Yam, M. A. Allison, R. Batdorf, H. Chen, H. Generazio, H. Singh, and S. Tucker, “The characteristics and emerging behaviors system of systems,” NECSI: Complex Physical, Biological and Social Systems Project 2004. [24] M. Maier, “Architecting principles of systems-of-systems,” presented at the 6th Ann. Int. Symp. Int. Council Syst. Eng., Boston, MA, 1996. [25] J. Boardman and B. Sauser, “System of systems: The meaning of OF,” presented at the IEEE Int. Syst. Syst. Conf., Los Angeles, CA, 2006. [26] IVHS America, IVHS Goals and Objectives IVHS America, System Architecture Committee. Washington, DC, 1991. [27] M. W. Maier, “On architecting and intelligent transport systems,” Joint Issue IEEE Trans. Aerosp. Electron. Syst./Syst. Eng., vol. 33, pp. 610–625, Apr. 1997. [28] USDOT, ITS Architecture Development Program Phase I: Summary Report U.S. Dept. Transportation, Nov. 1994. [29] USDOT. National Program Plan for ITS U.S. Department of Transportation, 1995. [30] in IEEE-USA Board of Directors, Operating Committee and Assembly, 2010 Meeting Summary, Atlanta, GA, 2010. [31] S. Ostry and R. Nelson, Techo-Nationalism and Techno-Globalism: Conflict and Cooperation The Brookings Institution. Washington, DC, 1995. [32] J. Howells and M. Wood, The Globalization of Production and Technology. London, U.K.: Belhaven, 1993.

[33] P. H. Appel, A Tour of the Research and Innovative Technology Administration, Implementing Transportation Solutions to Define a New Era, TR Research News 267 Mar.–Apr. 2010. [34] A. Saxenian, “The origins and dynamics of production networks in silicon valley,” Res. Policy, vol. 20, pp. 423–437, 1991. [35] B. Carlsson and R. Stankiewicz, “On the nature, function and composition of technological systems,” in Technological Systems and Economic Performance: The Case of Factory Automation, B. Carlsson, Ed. Boston/Dordrecht/London: Kluwer, 1995. [36] L. Soete, B. Verspagen, and B. Weel, “SoIs,” Working Paper Series, 2009-062 United Nations University, Maastricht Economic and Social Research and Training Centre on Innovation and Technology. Maastricht, The Netherlands, 2009. [37] C. DeBresson and D. Amesse, “Networks of innovations: A review and introduction to the issue,” Res. Policy, vol. 20, pp. 363–379, 1991. [38] B. Carlsson, S. Jacobsson, M. Holmen, and A. Rickne, “Innovation systems: Analytical and methodological issues,” Res. Policy, vol. 31, pp. 233–245, 2002. [39] H. W. Chesbrough, Open Innovation: The New Imperative for Creating and Profiting From Technology. Boston, MA: Harvard Business School Press, 2003. [40] A. Arundel, E. Lorenz, B. A. Lundvall, and A. Valeyre, “How Europe’s economies learn: A comparison of work organization and innovation mode for the EU-15,” Ind. Corp. Change, vol. 16, no. 6, pp. 1175–1210, 2007. [41] J. Bercovitz and M. Feldmann, “Entrepreneurial universities and technology transfer: A conceptual framework for understanding knowledge-based economic development,” J. Technol. Transf., vol. 31, no. 1, pp. 175–188, 2006. [42] A. Johnson and S. Jacobsson, “The emergence of a growth industry: A comparative analysis of the German, Dutch and Swedish wind turbine industries,” presented at the Schumpeter Conf., Manchester, U.K., 2001.

Ali Mostafavi is pursuing the Ph.D. degree and is a research assistant in the School of Civil Engineering at Purdue University, West Lafayette, IN. He is also pursuing the M.S. degree in industrial administration at the Krannert School of Management at Purdue University with concentrations on finance and innovation and technology management. His main research interests include sustainability and resilience in infrastructure systems, financial innovations in global infrastructure, innovation systems, innovation in infrastructure systems, and system of systems analysis and simulation.

Dulcy M. Abraham received the Ph.D. degree from the University of Maryland, College Park, in 1990. She is a Professor with the School of Civil Engineering, Purdue University, West Lafayette, IN. Her current research and teaching interests include infrastructure assessment and rehabilitation, value-added delivery of capital-intensive projects, project and program controls, innovative financing for infrastructure projects, mitigation and post-disaster recovery planning, construction safety and global issues in engineering and construction. Prof. Abraham served on the American Society of Civil Engineering (ASCE) Construction Research Council (CRC) Executive Committee from 2005–2008 (secretary, vice-chair and chair). Other national committee assignments include the ASCE Committee on Construction Equipment and Techniques Committee, the ASCE Water Infrastructure Security Enhancements (WISE) Committee, the ASTM Committee on Technology and Underground Utilities, and the WERF (Water Environment Research Foundation) Project Committee on the Examination of Innovative Methods Used in the Inspection of Wastewater Collection Systems.


Daniel DeLaurentis is an Associate Professor in the School of Aeronautics & Astronautics, Purdue University, West Lafayette, IN. He leads the System-ofSystems Laboratory (SoSL), which includes 12 graduate students and two Post-Doctoral researchers, and one Research Scientist. His primary research interests are in the areas of problem formulation, modeling and system analysis methods for aerospace systems and systems-of-systems (SoS). His research is conducted under grants from NASA, FAA, Navy, and the Missile Defense Agency. Dr. DeLaurentis is an Associate Fellow of the American Institute of Aeronautics and Astronautics and served as Chairman of the AIAA’s Air Transportation Systems (ATS) Technical Committee from 2008–2010. He is also Co-Chair of the System of Systems Technical Committee in the IEEE System, Man, and Cybernetics Society and is an Associate Editor for IEEE SYSTEMS JOURNAL.

265

Joe Sinfield is an Associate Professor in the School of Civil Engineering at Purdue University, West Lafayette, IN, and a Senior Partner at Innosight, LLC, an innovation consulting and investment firm headquartered in Watertown, MA. His research, teaching, and professional activities focus on the intersection of innovation, business, and technology. In professional practice, he has spent over a decade advising corporate leaders on issues of growth, technology investment, and innovation management. Through his research and teaching, he explores the principles of successful innovation in entrepreneurship and engineering business management, and applies them to develop novel sensing technologies for engineering challenges. His research has received both institutional and industrial sponsorship. Dr. Sinfield serves on the innovation advisory boards of multiple companies, he is the co-author of The Innovator’s Guide to Growth: Putting Disruptive Innovation to Work (Harvard Business School Press, 2008), and has published in a broad array of business, scientific, and engineering journals.