روش مدلسازی عامل تعاونی برای برنامه ریزی عملیات

A well designed computer-aided process planning (CAPP) system bridges the gap between CAD and CAM. A number of systems have recently been developed relying on a standalone expert system. However, because of over-complexity, many such systems cannot be effectively applied to industrial enterprises in practice. Moreover, the modern computer integrated manufacturing system (CIM) requires the CAPP system to be extendible and flexible for practical industrial applications. It is hardly possible to develop the extensive industrial CAPP system by using only one large expert system. To overcome these weaknesses, a new cooperative agent model is presented for process planning in this paper that satisfies five major requirements: Autonomy, Flexibility, Interoperability, Modularity and Scalability. In accordance with this framework proposed, a machining cooperative process planning system (Machining CoCAPP) is specifically developed for demonstration purpose. The system modelling, agent structure design, cooperation and coordination mechanism, and case study of the Machining CoCAPP are presented.

Process planning provides information to the shopfloor on how to produce the designed products. It addresses each part of the product separately and collectively. It defines the process, cost and production lead time under the constraints such as the designed geometry, material, quantity, machine and tooling availability, labour capacity and suitability, etc. In the past, process plans were often generated by human process planners who had plenty of manufacturing domain knowledge and worthy experience. In the recent decades, computer technologies have stimulated the advance toward computer-aided process planning (CAPP). Generally, there are two CAPP approaches: variant and generative. The variant approach is a data retrieval and editing method. Some standard or mature process plans are collected based on the group technology and stored in a database. When a new part is required to be produced, a similar process plan is retrieved from the database and edited to adjust it to suit the new part. The generative approach is a knowledge-based method which automatically generates a process plan according to the part's features and manufacturing requirements. The success of the variant approach depends on the group technology, planner's experience and a sufficient collection of standard or mature process plans. This method is especially suitable for companies with few product families and a large number of parts per family. Most earlier CAPP tools can be categorized under the variant process planning approach [1]. Typical examples are CAPP [2], MIPLAN [3], etc. The generative process planning approach has attracted more attention in recent years. It offers a potential of producing an optimal plan. Typical examples are APPAS [2]EXCAP [4], KRONOS [5], XCUT [6], QTC (Quick turnaround cell) [2], PART [7], OOPPS (object-oriented process planning system) [8], MePlans [9], COMPLAN Process Planner (CPP) [10], etc. Generative process planning systems are mostly oriented towards the needs of large companies and research organizations, especially those which have a number of products in small lot sizes. However, there is still difficulty in developing a truly generative process planning system which can meet industrial needs and provide an appropriate and compatible generic framework, knowledge representation method, and inference mechanism. Cooperative agent systems attempt to distribute activities to multiple specialized problem solvers and to coordinate them to solve complex problems 11, 12, 13 and 14. A cooperative agent system consists of many individual agents with cooperation engines. Each agent which has its own knowledge base and inference engine is responsible for a specific task. It provides a cooperation interface to communicate with other agents in the cooperative environment. A different language and different knowledge representation may be employed by each agent which may well be located in a different machine. Such a system organization provides an integration environment of heterogeneous and scalable architecture suitable to solving different problems.

The cooperative agent model for CAPP was introduced in this paper. The model makes use of intelligent agents and tries to satisfy five major requirements simultaneously: Autonomy, Flexibility, Interoperability, Modularity, and Scalability. An experimental Machining CoCAPP system has been developed by using the proposed model. The developed CoCAPP system is different from other CAPP systems available; it utilizes cooperative and coordination mechanisms built into distributed agents with their own expert systems. Each agent in this system deals with a relatively independent domain of process planning. This is in sharp contrast to other CAPP systems utilizing a single standalone expert system to perform the entire process planning. The system is hence flexible and upgradable. This feature is especially useful as the change of process planning methods or revision due to technology advances is increasingly more common and frequent. In the paper, a typical mechanical component is considered to test the performance of the Machining CoCAPP system. The experimental results show that the system can effectively deal with the process planning problems. It can generate process plans according to the product design data and available manufacturing resources. The system has met the proposed design requirements. In particular, the CoCAPP system has the following characteristics. • Its P-agents can be added and deleted at any time without affecting system operation, and can be individually updated without affecting the others, thus reflecting the modularity and flexibility features embodied in the system. • Each P-agent only generates a partial solution related to its knowledge; therefore, a complex problem can be decomposed into many simpler sub-problems on a modular basis. • The whole solution to a problem is obtained by integrating each P-agent's proposal together with other proposals. • Each P-agent can be an individually developed expert system or an analytical program with cooperation knowledge included according to demand. This autonomous feature greatly simplifies the implementation of the CAPP system. This paper presents the method to model the process planning agent (P-agent) by using intelligent agent technology. According to this model, each agent is interoperable and not confined to any machine platform. The model for conflict resolution strategy is developed to suit the CoCAPP system. In addition, the process-planning knowledge base is divided into multiple knowledge bases which are independently established. This has greatly reduced the search space of each inference engine. Based on the proposed CoCAPP system, the optimization of process plans would be feasible and easier to obtain. The experimental results have shown that the CoCAPP framework can be easily integrated into the concurrent engineering environment to implement integrated product design; it can deal with unreasonable part designs. The proposed CoCAPP framework opens up a new approach to the CAPP development. It provides an open framework. It is very suitable for distributed CAPP system development. Further investigations should focus on the improvement and extension of the system. Currently, the experimental Machining CoCAPP system only includes three P-agents. Other P-agents such as feature recognition, etc. may be added to the system in the work of future development.