Operations management research in process industries
综述了流程工业运营管理研究的现状，指出该领域研究不足，并强调需整合实证、概念和建模三个视角，以推动该领域发展。
The process industries cover a wide range of businesses, ranging from continuous facilities in the petrochemical industry, to large batch manufacturing in steel production and glass manufacturing, to small batch manufacturing in the food and pharmaceutical industry. While the process industry realizes a considerable portion of GDP in many countries, operations management (OM) research has traditionally paid very little attention to this large group of industries. The first significant contribution that recognized the field of process industries within operations management has been the pioneering work of Taylor and his research group (e.g., Taylor, 1979; Taylor et al., 1981). Their main contribution was to document the differences between discrete and process industries as illustrated in the often-quoted article “Why the process industries are different”. Along with stressing the different nature of process industries, they developed methods and tools to deal with these differences in planning systems. Specifically, the inadequacy of standard MRP-systems for capacity planning as one of the main specific elements in the process industries has been emphasized, primarily due to the APICS context within which much of their research work was conducted and published. Along the same line of thought the practitioner work of Van Rijn et al. (1993) can be positioned. They describe an exhaustive list of characteristics of large batch and small batch process industries (which they denote as “semi-process industry”), which is a useful base for further studying semi-process industries. These authors use it to derive consequences for the introduction of MRP-systems in those industries, and provide a toolbox for doing so. In the same line of research, Van Donk (2001) derives a list of specific characteristics of the food processing industries. Starting from the early nineties, more attention has been paid to differences within process industries rather than the differences between process industries and discrete manufacturing. Starting from a general list of characteristics, Fransoo and Rutten (1994) distinguish between two basically different types: flow process industries and batch process industries. They propose that all process industries can be placed on a continuum between these two extremes. Moreover, they derive on the basis of these differences consequences for planning and control. Their work makes clear that much of the earlier work that positioned process industries as opposed to discrete manufacturing, in fact, implicitly focuses on large batch manufacturing and fail to address the highly specialized batch process industries. While the work of Fransoo and Rutten (1994) lacks a documented empirical approach, the studies by Dennis and Meredith (2000a, 2000b) continue with this perspective of differences within the process industries. Their empirical study among 19 firms in the process industry demonstrates that within the process industries well-defined and different types can be distinguished and that different production and operations management systems are used. They also list some of the characteristics of these types. The above mentioned papers add to our knowledge from an empirical angle and from a conceptual angle by conceptually mapping its characteristics. As a result the field of process industries has been mapped, differences have been identified, and it has been demonstrated that a considerable variety exists within what is commonly recognized as “the process industry”. In order to assess the current status of operations management research in process industries, we need a third angle that is strongly related to what can be characterized as traditional operations management theory. With that we refer to the theoretical, analytical, and modelling work within operations management: also labelled as production planning and control (PPC) models. Within this area a lot of work has been conducted to derive decisions rules for specific sets of characteristics under certain conditions. However, much of the work proposing models lacks specific knowledge of the process industry domain, enforcing that many of the characteristics are either assumed too general (“the process industry”) or not addressed specifically. We argue that each of the three areas or angles described so far, is useful for assessing the current status of research in the field. Moreover, the different areas should be related to each other and use each others results in developing the field as a whole. We think that the absence of these relationships is the main problem. The relationship between reference models for production planning and control, empirical studies describing the field of process industries, and studies describing specific characteristics of process industries as distinct from discrete manufacturing has been depicted in Fig. 1. Main building blocks for assessing research in the process industries. Within each of the fields of Fig. 1, important results have been established, but the relationship between them have not been addressed. We have described above that the field has been documented empirically in the studies by Dennis and Meredith (2000a, 2000b). While their approach learns us the characteristics of different types of process industries, and these types are consistent and further detail out the propositions by Fransoo and Rutten (1994), they do not make an explicit link to either the available models for production planning and control decisions or to the now omnipresent lists of characteristics of process industries (of which Table 1 in Dennis and Meredith, 2004a provides a comprehensive overview). The work by Dennis and Meredith (2000b) can be seen as a sound basis, following which further research is needed to both: (1) extend the empirical basis of the seven types by further field studies and (2) to make more explicit which of the often-cited characteristics are dominantly present in the types identified by them. Within the modelling literature, a wide variety of studies can be recognized that have or may have substantial impact in the process industries. Much of the lot sizing literature, for instance, is based on the single machine, limited number of products assumption that is present in many large batch manufacturing plants in the chemical industry. Since an explicit link between many of these basic models, which we denote as reference models for production planning and control, and the empirical literature in process industries is non-existent, it is oftentimes unclear whether claimed lacks of knowledge in the process industry literature have been addressed elsewhere. Likewise, in many occasions specific characteristics of process industries may not have been taken into account in the reference models. This issue has also been noted by Flapper et al. (2002) who provide an overview of research issues in remanufacturing in process industries. They make an explicit link between modelling work and perceived characteristics of manufacturing in the process industries, with a specific emphasis on remanufacturing. Here we need on the one hand research to develop specific reference models related to specific characteristics of the process industry and on the other hand empirical work to find specific production planning models in the process industries (developing further the work of Dennis and Meredith, 2000a). The objective of this special issue has been to demonstrate good examples of operations management research in the process industries, which do take at least two of the perspectives of Fig. 1 into account. We will shortly introduce each paper and relate its contribution to the above figure. Teunter and Flapper (A comparison of bottling strategies in the pharmaceutical industry) relate a specific characteristic of the pharmaceutical industry to building a decision model for planning and control of inventories and the location of intermediate product. The model developed can be seen as a reference model for processes where quality of products can only be determined after a certain (quarantine) period of time and one has to decide to pack the product or to store it unpacked for that period. Food and pharmaceuticals are typical application areas for the decision models derived, but still more work has to be done to apply the results in the field. The paper by Cooke and Rohleder (Inventory evaluation and product slate management in large-scale continuous process industries) is based on a case study. The main contribution here is that results form the PPC model area are applied and adapted to the typical situation of many chemical industries: the effects of off-grade production and production rate differences when changing and starting up production. Here the contribution is mainly to a specific chemical industry, but the paper also adds to our understanding of reference models for the process industry. Ketokivi and Jokinen (Strategy, uncertainty and the focused factory in international process manufacturing) combine a number of different methods to collect interview data, production figures and site visits in order to address a well-known concept in production management: the focused factory. The paper explores the issue which circumstances favour focused factories and which do not. As such the paper contributes to our knowledge in the differences between seemingly similar factories. Moreover, it shows that process characteristics are not the only determining factor for the choice of manufacturing strategy, which was known and proved for discrete industries, but is less evident for process industries. The paper thus contributes to each of the three relationships. French and LaForge (Closed-loop supply chains in process industries; an empirical study of re-use issues) present a nice example of survey research. Starting from the “commonly held believe” of typical process industry characteristics, this paper concludes that there is a lot of variety in re-use strategies in process industries. The paper adds to our knowledge of the empirical differences in process industries and the consequences for production planning. As such it can help us to develop better models and aids for planning such closed loop supply chains. Van Wezel, Van Donk and Gaalman1 (The planning flexibility bottleneck in food processing industries) reflect on their previous work in small and medium sized food processing industries. Whereas the common believe is that capacity is the main bottleneck in planning, the authors show that organisational arangements often are a barrier in flexible response to customer's requirements. This paper helps us to better understand organisational reality and can be of use in developing and implementing reference models and stresses the importance of organisational characteristics in process industries. The field of process industries is quite active within the operations management research community, since we received 29 papers for this special issue, out of which we finally accepted five papers. We would like to thank the many reviewers that contributed to this special issue. Furthermore, we would like to thank the Editor-in-Chief, Robert Handfield, for giving us the opportunity to edit this Special Issue and for his support and help in the editorial process.