基于正/逆向网络SimEvents仿真的PERT网络分析

PERT Network Analysis Based on Forward/Backward Network Simulation in SimEvents

针对工序持续时间服从任意分布的PERT网络,基于离散事件系统(Discrete Events System,DES)Monte-Carlo仿真的思路,在数学证明的基础上,将逆向网络仿真与正向网络仿真相结合,以获取以往DES仿真较难获取的与逆向回馈计算有关的时间参数;探索性地选择用MATLAB新增的离散事件仿真工具箱SimEvents对正、逆向网络中的节点、弧线和网络进行仿真建模,使模型具有直观、建模简单和子系统可复用等特点。以仿真结果为基础,可获取丰富的有关网络、工序、节点、时差的各类信息,以增强对网络,尤其是大型、复杂、多层次网络运作细节的了解和对任务过程的控制能力。

Based on mathematical proof, a new Monte-Carlo simulation method based on Discrete Event Systems（DES） was developed to analyze PERT network. The new method combined the DES simulation of PERT network with the DES simulation of its backward one. Through it, a wealth of time parameters could be obtained, including the ones which must be calculated through cumbersome forward and reverse recursive computation by conventional analytical methods. MATLAB’s new discrete event simulation toolbox SimEvents was used to build the simulation models, through nodes, arcs to networks. Its advantages of subsystem package and reusability made the model very simple and intuitive. If combined with MATLAB＂s powerful ability of calculation and graphical representation, this method could enhance analyzers’ understanding of the operational details, and therefore strengthen their ability of controlling the task process, especially the great, complex and multi-level one.