利用启发性知识降低交叉算子破坏性的研究

在随机搜索的过程中,交叉算子被认为是演化算法的主要动力.然而,在搜索更好的候选解的时候,如果交叉点选择不当就会破坏一个已有的较好的候选解,使得交叉后生成的新候选解比原有的候选解更差.该文提出应用与问题相关的领域知识构建更有建设性的交叉算子,并应用于旅行商问题(TSP).实验结果表明,基于具有建设性的交叉算子的混合演化算法具有更好的性能.

Approach towards a Less Destructive Crossover Operator with Heuristic Knowledge

During the process of stochastic search,crossover is considered to be the major driving force behind evolutionary algorithms.However, when searching for good scheme,a good solution can be destroyed by an inappropriate choice of crossover points,so many attempts have been made to make it a more constructive crossover.In this paper,we introduce a more constructive crossover operator utilizing the problem-specific knowledge,and applying this method for the traveling salesman problem(TSP).The experimental results show that the hybrid evolutionary algorithm with a more constructive crossover operator has better performances than other methods.