The Research On Modeling And Scheduling Of FMS Based On Petri Net

The design and implementation of flexible manufacturing systems (FMS) are very complicated. Because FMS are costly, risky and technology-intensive, we should find a proper method to model and analyze the FMS before they are put into practice. By modeling and analyzing FMS, we can find the systems' faults and can diagnose and recover them in the process of real-time control more easily. Besides, the scheduling and controlling of FMS are also key problems in FMS. Based on the model of FMS, we can simulate the systems according to some specific rules and input the information about devices and rules to the system. We can also evaluate the data about the system performance for correcting systems or finding an optimal scheduling path. Now, because Petri nets can describe the share, conflict, mutex, concurrence and uncertainty of resources and their model can be analyzed quantitatively and qualitatively, several kinds of Petri nets have become the important methods of modeling and scheduling FMS.This dissertation concentrates on modeling and scheduling FMS based on Petri nets and proposes several simple, effective and practical heuristic algorithms to solve some classical FMS scheduling problems with NP difficulty. Generally, the main achievements in this paper are as follows:(1) Develop a software to modeling and scheduling FMS. This software provides simulation environment for testing our methods. It has implemented the static and dynamic traits of Petri nets and can use the input files of FMS' Petri net models easily. Besides, it also can generate specified number of test models accord with some specified conditions and we can use the software to test different algorithms by replacing the function of schedule. Using the software, the user can concentrates on the problems they want to resolve.(2) A problem observed in algorithm L1 is that this scheduling algorithm may not be admissible as stated in its paper. This problem was also recognized by some people who devised a remedy for this problem. But their method still can not guarantee that it does not reject markings that may lead to the optimum solutions. This paper presents an improved algorithm which thoroughly corrects the defect in their methods and can obtain optimum solutions with an admissible heuristic function. To speed up the search process, another algorithm is also proposed which can still conserve the admissibility of the algorithm but at a reduced search...