عنوان مقاله
توسعه و اجرای یک فرایند dispersed decision : یک نمونه از زمان‌بندی FMS
فهرست مطالب
چکیده
مقدمه
ابزار تحقیقاتی
روندهای ارتباطی و محیط عملیاتی
سیستم پردازش کننده برنامه
نتیجه گیری
بخشی از مقاله
وقتی که مراحل کامل شد، هر ایستگاهlocal از میزان مزایده های برنده شده باخبر می‌گردد، و متوجه می‌شود که کدام مزایده ها را برده و کدام مزایده ها را از دست داده است. به برندگان یک زمان مشخص داده می‌شود تا تصمیمات آزمایشی شان را بر روی درآمدهای مزایده تغییردهند. زمانی که این مرحله کامل شود، سیستم بازه زمانی دوم در حلقة مزایده را شروع می‌کند. زمانی که آخرین مزایده در بازه زمانی مشخص کامل شد، مزایده کننده "پایان مزایده" را اعلام می‌کند تا همه فرایندها تعطیل شود.
کلمات کلیدی:
Development and implementation of a dispersed decision process: an FMS scheduling example Shung-Kuang Kung* and James R. Marsden’ *Department of Management Information Systems, College of Business, Chung-Yuan Christian University, Chung-Li, Taiwan ‘Department of Operations and Information Management, 368 Fairfield Road, U-41 IM, School of Business Administration, University of Connecticut, Storrs, CT 06269-2041, USA An automated process was constructed which provides a platform for conducting analyses in general distributed multi-participant decision making environments, including the special case of FMS scheduling. The detailed example illustrates the use of the tool for analysing distributed decision making in a multi-machine, multi-task FMS setting. The example relates a controlled laboratory experiment using human ‘operators’ interacting in an electronic auction. The generality of our automated process supports this type of experimentation, as well as complex simulations using computerized expert systems as ‘participants’, simulations that enable us to perform detailed comparisons between the performance of the distributed process and existing centralized FMS scheduling heuristics. Keywords: flexible manufacturing systems, distributed decision making, laboratory experiment Introduction Production decision processes and mechanisms may conjure up physical relations and tangible products. But what is there that is truly unique about these settings? Are they truly singular, or are they set apart because of restrictive definitions that we choose to characterize them? Consider flexible manufacturing systems (FMS). One commonly cited definition is that of Kusiak’: a set of machine tools linked by a material handling system, all controlled by a computer system. Suppose, instead, that we adopt a much broader approach, defining an FMS as a production system where one or more machines are capable of performing more than one task (where suitably broad definitions of ‘machine’ and ‘task’ are intended). We note that this definition is sufficiently expansive to include a principal (major contractor) dealing with numerous potential subcontractors. In what follows, we demonstrate how, using such a broad definitional approach, we are able to construct an automated process which provides a platform for conducting analyses in general distributed multiparticipant decision making environments, including the special case of FMS scheduling. Our detailed example, provided below, illustrates the use of the tool for analysing distributed decision making in a multimachine, multi-task FMS setting. The example relates a controlled laboratory experiment using human ‘operators’ interacting in an electronic auction. The generality of our automated process supports this type of experimentation as well as complex simulations using computerized expert systems as ‘participants’. Though much of our motivation for developing the tool rested with our desire to be able to investigate thoroughly alternatives to centralized scheduling in existing and likely future FMS environments, we quickly realized that the approach held possibilities for investigating the performance of more general distributed decision making environments. Whether dealing with extremely complex (NP-complete) FMS scheduling problems or with less demanding assignment problems, a key determination must be the performance comparison of centralized versus decentralized decision making processes. As detailed below, our automated process provides a convenient tool for analysing the performance of distributed processes. The next section sets forth the specifics of the general environment we model and the automated process we have developed to perform analyses within this general environment. Two FMS example experiments we conducted are then detailed and the experimental outcomes summarized. Finally, our concluding remarks and suggestions are provided. Research tool The research tool we implement is a flexible operationalization of the approach set out in Shaw and Whinston2. These authors ‘followed the distributed
