论文部分内容阅读
进化计算已经被成功地用于模糊系统自动生成.但是当输入变量增加时,一个个体对应整个模糊系统的编码方式往往会因编码太长而降低进化的效率.但每个个体代表一条规则又会给适应度评价带来困难.本文提出了一种把合作式协同进化算法用于模糊系统自动生成的新方法.每个个体代表一条或几条规则组成的子模糊系统,把所有个体分为一些子种群,这些子种群进行合作式协同进化,引入一个自适应机制动态调整种群个数,最后从每个子种群中选出最佳个体构成完整的模糊系统.实验结果显示该算法提高了进化效率.最后对个体定义等相关问题进行了讨论.
Evolutionary computation has been successfully used to automatically generate fuzzy systems, but when the input variables increase, the coding of an individual corresponding to the entire fuzzy system tends to reduce the efficiency of evolution because the coding is too long, but each individual represents a rule Which brings difficulties to the fitness evaluation.This paper presents a new method of using co-evolution coevolutionary algorithm for the automatic generation of fuzzy systems.Each individual represents a sub-fuzzy system composed of one or several rules, which divides all the individuals into some Subpopulation and the co-evolution of these sub-populations, an adaptive mechanism is introduced to dynamically adjust the population number, and finally the best individuals from each sub-population are selected to form a complete fuzzy system.The experimental results show that the algorithm improves the evolutionary efficiency. Finally, the individual definition and other related issues were discussed.