Granularity Fitness Landscape and Directed Mutation in Evolutionary Algorithm
This paper introduces the granular computing thought and methods into evolutional- algorithm. Take the solving process of the evolutionary algorithm as an information system; adopt the equivalent relation to divide the solution space and objective function, form subspace granule or objective function granule. On this basis, we proposed the conception of granularity fitness landscape. We used the methods of binary granular computing to find the relation between subspace granule and objective function granule, expressed the dependency of objective function to the individuals at macroscopic. Took this as foundation to carry out the mutation operator what was direct by granular fitness landscape, this operator can improve the convergence rate of the evolutionary algorithm remarkable.
evolutionary algorithm granular computing granularity fitness landscape directive mutation operator
Yan Gao-wei XIE Gang Zheng Zhong
College of Information Engineering Taiyuan University of Technology Taiyuan, China
国际会议
海口
英文
99-104
2011-02-22（万方平台首次上网日期，不代表论文的发表时间）