تعیین معیار و برنامه ریزی تولید و کنترل مبتنی بر چند عامل قوی
|کد مقاله||سال انتشار||مقاله انگلیسی||ترجمه فارسی||تعداد کلمات|
|5592||2003||14 صفحه PDF||سفارش دهید||7569 کلمه|
Publisher : Elsevier - Science Direct (الزویر - ساینس دایرکت)
Journal : Engineering Applications of Artificial Intelligence, Volume 16, Issue 4, June 2003, Pages 307–320
Multi-agent systems (MAS) offer new perspectives compared to conventional, centrally organised architectures in the scope of production planning and control. They are expected to be more flexible and robust while dealing with a turbulent production environment and disturbances. In this paper, an MAS is developed and compared to an Operations Research Job-Shop algorithm using a simulation-based benchmarking scenario. Environmental constraints for a successful application of MAS are identified and classified. Furthermore, the topic of MAS robustness is addressed by applying database technologies on the basis of transactions.
Companies nowadays have to face a global market characterised by numerous competitors, a steadily increasing complexity of business processes and a highly turbulent production environment. Consequently, manufacturing systems have to provide the flexibility and reliability that is required to stay competitive. Decentralised planning and controlling approaches offer interesting perspectives compared to conventional centralised architectures. In the scope of production planning and control (PPC), multi-agent systems (MAS) are expected to be more flexible than centrally organised systems. Nevertheless, they lack of reliability and robustness that is necessary for an industrial deployment. To prove or disprove the thesis of MAS being more flexible and thus being able to increase the planning quality for well-defined shop floor scenarios, a simulation-based benchmarking platform on the basis of a real test case scenario was developed at the University of Karlsruhe in the scope of the Karlsruhe Robust Agent SHell (KRASH) project. A performance measurement system is included to provide not only qualitative, but also quantitative results. The platform is used to compare existing PPC approaches based on Operations Research (OR) algorithms with decentralised MAS approaches. Furthermore, different scenarios can be simulated with various levels of complexity. This makes it possible to set up a map that identifies application scenarios, where MAS provide a real benefit to potential industrial users. In the next step, abstract rules can be extracted from these results to gather further knowledge about the preferences of MAS. Besides the quality of the planning results, robustness is a very important aspect of a manufacturing system, especially since the focus of the project is set upon handling disturbances like machine troubles or tardiness caused by external suppliers. On the shop floor, reliability may be guaranteed by sophisticated planning algorithms. On the other hand, the software implementation of the MAS has to be robust, too. Due to the distributed architecture consisting of autonomous and intelligent entities, MAS are more error-prone compared to central approaches. Thus, special attention has to be paid to technical robustness issues. Robustness and reliability are common features of modern database systems. Consequently, database technologies are used to provide services that guarantee the robust execution of agent tasks. The implementation of robust MAS is simplified by defining a framework for the transaction-based execution of agent tasks. Local and dispersed agent plans are executed in a robust way by using a transaction service. Its performance and scalability was evaluated by using simulation technologies. In Section 3, the simulation-based benchmarking platform is described. The developed MAS approach is presented in Section 4. Section 5 shows the results of the comparison of the centralised and decentralised planning approaches and draws conclusions of the results. The transaction-based robustness service and simulation results are introduced in Section 6. Section 7 summarises this paper.
نتیجه گیری انگلیسی
The presented approach has a mentionable disadvantage regarding the usability. Agents have to represent their plans and potential compensation activities within the format of open-nested transactions, i.e. the robustness of the MAS is achieved by demanding an additional implementation effort from the developer, especially if distributed agent plans are considered. This motivates the request for a more “natural” and “transparent” integration of the robustness service into the MAS implementation process. As a consequence, in our current research, a different approach is chosen, based on distributed ACID-transactions, which are identified with conversations between negotiating planning agents. The related advantage is a natural and seamless integration of the mechanism into standardised MAS platforms like FIPA-OS. The system architecture is depicted in Fig. 18.In the new approach, the original User Tasks, that handle the planning conversations in each involved agent, are embedded into transactional tasks (TA Tasks). They later automatically handle all necessary additional duties caused by the use of the robustness service for their embedded User Tasks. As a result, the realised approach only interferes with the communication layer and does not affect processes within the agents themselves. On the other hand, distributed ACID-transactions do not make use of the complete semantics of the agent conversation, i.e. they are probably too restrictive. This may lead to a decrease in performance or cascading recoveries. These considerations are focused by our current research activities. The final goal comprises a flexible, conversation-based transaction model that also takes into account the internal agent states. A combined mechanism, consisting of the first and second approach described above and including knowledge about the semantics of the messages inside a conversation, may be a promising new approach.