遗传算子概率在约束优化问题上的自适应探讨

Discussion on the Adaption of Genetic Operator Probabilities for Constrained Optimization Problem

提出两种算子概率自适应方案:基于连续两代间解的收敛状况变化的群体性方案、基于群体代间及当代表现的个性化方案.在带转包受限经济批量问题上,进行了变异概率调整方法的仿真试验,与文献中的其它变异概率调整方案进行了比较.结合试验结果和文献中的相关研究,指出在群体中存在不可行个体情况下,算子概率自适应的调整机理及理论分析仍是一个有待解决的问题.

Two adaptive schemes of genetic operator probabilities are proposed,one is based on the change ratio of the convergence situation of the solution in the continuous two generations,the other on population performance in the present generation and continuous two ones. For the economical lot sizing problem with outsourcing,the simulation examination with several adjustment schemes of mutation probability is made,their test results are analyzed. From the test results and relevant research in the literature,we indicate that adaptive mechanism of genetic operator probabilities as well as its theoretical analysis is still open when there are infeasible individuals in the GA population.