Modeling, Analysis And Hardware Implementation Of Discrete Event Systems Based On Petri Nets

Discrete event system(DES)has the characteristics of discrete state space and piecewise stable trajectory.It has many manifestations in manufacturing,transportation,robotics and other fields.The research and application of DES is the inevitable requirement of low energy consumption and high efficiency in social production.Petri nets can describe concurrent,mutually exclusive,conflicting and sharing behaviors of system resources concisely and intuitively,and can conduct qualitative and quantitative research on these behaviors.Therefore,using Petri nets to model,analyze and schedule discrete event systems has become one of the mainstream methods in the academic community.However,in the existing theory,Petri net face the problem of state explosion,that is,the set of reachable markings of Petri nets increases exponentially with the increase of net size.When dealing with large-scale systems,traditional Petri net analysis methods have some problems,such as too long computing time or memory overflow.This paper mainly studies the modeling and analysis methods of discrete event systems based on Petri nets.Aiming at the problem of state explosion in large-scale systems,several optimization methods are proposed to speed up the analysis and reduce the system model.In addition,the efficient modeling ability of Petri nets is applied to practical industrial systems.The main contents and innovations of this paper are as follows:(1)Aiming at the problem that the existing Petri net analysis software INA solves the reachable state set slowly,a parallel method of generating reachable markings set based on GPU and CPU is proposed.This method makes full use of GPU's efficient cooperative computing ability and expands the successor state in parallel.Moreover,the storage structure of reachable markings in memory is improved,which greatly speeds up the detection of duplicate markings.The experimental results show that the method of generating reachable markings based on GPU + CPU can improve the computing speed of large-scale Petri net model by hundreds of times compared with INA software.(2)Aiming at the problem that the number of BDD nodes is very sensitive to the sequence of variables when using BDD to analyze Petri nets,two methods of generating the sequence of variables in the symbolic analysis of Petri nets are proposed.The first method generates a heuristic sequence of BDD variables based on the structure information of the Petri net model and the correlation degree between the places.The second method firstly constructs Petri net causality diagram,and then find a sequence of variables that maximize causality by the solution for TSP problem.Experiments show that these two methods can get a better order of variables in a very short time,effectively alleviating the problem of state explosion in Petri net model analysis.(3)Aiming at the avalanche effect when Petri nets are converted to PLC programming language in the process of hardware implementation of the model,a method is proposed to prevent the avalanche effect,that is,first detecting the enabling conditions of all transitions in advance,and then firing the enabled transitions.This method also sets different priorities for the transition with conflict structure,the transition with higher priority is transmitted first,which avoids the logical ambiguity caused by the conflict structure of Petri net.