The evolutionary algorithms of self-adaptive crossover and mutation

高汉平,康立山,杨族桥,肖小红

DOI: https://doi.org/10.3969/j.issn.1003-8078.2003.03.020

2003-01-01

Abstract:In order to avoid increase of interation count and local opti mi zation solution resulted from the improper selection of crossovers and mutations during the evolution, we propose a new method with self-adaptive crossovers an d mutations which employs an improved strategy of the multi-parent crossovers a nd the non-uniform mutations. Experimental results show that the method not onl y avoids prematurity but quickly finds optimal solution with stable performance and high search efficiency.