期刊论文

An Effective Way in Parameters Optimization of Fuzzy Petri Nets

中文

电脑知识与技术

1009-3044

2007

4

19

208-210

10.3969/j.issn.1009-3044.2007.19.106

本校为第一机构

本文根据模糊Petri网运行时连续性的特点,以蚁群算法为基础,提出了一种带交叉、变异因子的模糊Petri网参数值寻优的有效方法。该方法先用分层思想建立无环路的FPN模型,然后把它转化为类前向神经网络以确定输入—输出层关系,并将网络中各参数的值域均分为10等份,在图形中用10个城市来表示,再用蚁群算法进行路径的寻优,寻优后,在蚁群选定的值域上产生具体的分量解,最后由误差函数确定是否需要重新寻优。仿真程序实例表明经蚁群优化算法训练出的参数正确率较高,且不依赖于经验输入数据。

This article devised an effctive way to optimize the parameters in fuzzy petri nets with intersectant genes and varietal genes.This technique used the thought of delamination to found a modle of FNP without loop,and then, transformed the modle to aquasi-feedforward neural network—confirmed the connection between input-stratum output-stratum, contemporary,we shared the parameters in nets and used the ten cities to express in graph,and to optimize the routes by ant colony algorithm.After this, at the value range of ant selected it could produced...