摘要
提出了一种新的基于N进制分部编码算子的遗传算法.该编码算子首先将每个基因值用N进制的浮点数表示,然后将其分为整数部分和小数部分,分别重新编码组成染色体;相应的选择、交叉、变异算子采用符号编码的思想,充分利用N进制浮点数的特点进行设计.在遗传算法开始阶段,该编码算子进行整数部分和小数部分的遗传操作,使得遗传算法在早期具有很强的全局搜索能力,避免陷入局部极值;在后期进行小数部分的遗传操作,使得遗传在后期具有很强的局部搜索能力,能够很快地搜索到全局极值.通过理论分析,证明了N进制分部编码算子与传统的浮点数编码和二进制编码算子相比具有优越性,并通过典型函数的仿真进行了验证.*
A new genetic algorithm based on N-decimal system parted coding operator is presented, The codingoperator expresses the genes with N-decimal system floating numbers, and then the genes are divided into integer part and decimal part. The two parts are encoded into chromosomes respectively, The corresponding selection, crossover and mutation operators are designed with the theory of symbol encoding and the properties of N-decimal system floating numbers. In the early stage of the genetic algorithms, the three genetic mechanisms are used both in integer part and decimal part, so that the genetic algorithms have stronger global search ability, keep the population diversity efficiently and avoid falling into local extremum. In the later stage of the genetic algorithms, the three genetic operators are used in decimal part, so the genetic algorithms have stronger local search ability and fast convergence ability. Theoretical analysis shows that the N-decimal system parted coding mechanism is more efficient than the floating encoding method and the binary encoding method, Simulations on some representative functions are given to validate the results and theory.
出处
《信息与控制》
CSCD
北大核心
2006年第5期624-628,633,共6页
Information and Control
基金
国家自然科学基金资助项目(30371362)
关键词
编码算子
遗传算法
优化
coding operator
genetic algorithm
optimization
