一种防止浮点遗传算法早熟收敛的父代选择策略.pdfVIP

第17 卷第5 期 2002 年9 月 Vol. 17 No.5 Control and Decision 　Sept . 2002 : 1001-0920(2002) 05-0625-04 谢晓锋, 张文俊, 杨之廉 ( , 100084) : 通过对浮点遗传算法早熟收敛现象的分析, 提出了 一种新的父 选择策略, 即使用当前 的子 个体作为下 的父 个体, 可使交叉算子持续地探索和开发新空间。引入对 个体的 数保护策略, 即 在它发生变异前保证有足够的演化, 可以避免对新空间不成熟的开发。通过与其它父 选择策略的对 比, 并通过实验和GENOCOP 系统比较, 表明本方法能得到较好的结果。 : 遗传算法; 早熟收敛; 选择策略 : TP 301. 6: A parents selection strategy fighting premature convergence in floating genetic algorithms X I E X iao-f eng , ZH A N G Wen-j un, YA N G Zhi-lian (Institute of Microelectronics T singhua University Beijing, , 100084, China) Abstract: With the analysis of the premature convergence of floating genetic algorithms (FG s) , a new selection strategy is proposed by using the children individuals of current generation as the parent of next generation, which makes the crossover operator to explore and exploit in new spaces successively. protection strategy for the evolved generations for individuals is introduced to guarantee enough evolution before the mutation occur, which avoids immature exploitation for new spaces. Compared with the results of GENOCOP system by experiment this selection strategy provids better solutions, . Key words: genetic algorithms; premature convergence; selection strategy 1 引　　言 , [ 1] ( G s) Holland , ,, [5] , G , G , , , ,