Study And Implementation Of The Modeling For PPCES Based On Extended Petri Net

The single piece and little batch production is a kind of general production type in discrete manufacture. The high -specialties, variety, low repetition, long cycle of the production is its feature, and the production load is dynamic. The schedule of this kind of production can be formulated as assigning N parts to M machines. To find the optimal solution is an NP-complete problem in theory. So far, there is no efficient algorithm to solve the problem.Petri net is a kind of graphic modeling tool with strong descriptive ability. It cannot only describe the dynamic procedure of discrete events but also the parallel, synchronous, conflicting and consequential relations. So it is especially suitable to be applied in the modeling and performance analyzing such production systems. It is a comparative effective method to establish model and then analyze it.But if Petri net is directly adopted to model the real problems, too many nodes will be produced, which leads to the exponential increase of the search space scale, so the model is difficult to be analyzed with the usual methods. Although we can reduce the complexity of the model through simplifying rules and methods, we have to repeat the process when there is any little change in the system model, because the original analyzing result can not reflect the new feature of the changed system any more. So an idea of combining the analyzing method of Petri net and the expert system is brought forward to reduce the scale of the research space.First, a kind of extended Petri net is defined. The timed and colored elements are added. It can reveal the relations among procedures and the demand of the procedures for key machines as well, so the descriptiveability is greatly enhanced. Second the reasoning load is much reduced by adding running rules for the model so that the model has the mechanism to deal with parallel and competitive phenomenon and the automotive reasoning ability. In addition, a method to keep the consistency of the knowledge database by taking use of Petri net's structure is put forward. The whole system has been successfully developed with Visual C++ under Windows.