一种改进的遗传算法
|
摘要
采用实数编码的遗传算法 ,在基于适应值比例和最优保留策略结合的选择、数值交叉及一致变异的基础上对算法作了两方面的改进 ,即对交叉操作采用剔除无效个体和保留父代进入选择 ,在算法出现未成熟收敛的趋势时插入混沌序列 ,增加群体多样性 ,以判断算法搜索特性。对函数优化仿真结果显示 ,新方法提高了搜索精度 ,较好地克服了未成熟收敛现象 ,取得了较满意的优化效果。
Based on the combination of fitness property selection and elitist model, numerical crossover and uniform mutation, two methods have been put forward to improve the efficiency of real-coded genetic algorithms. The first method is eliminating the inefficient individuals in crossover operation and reserving the parents for participating selection. The second method is inserting chaos serials into the population when a trend of immature convergence appears. Simulation results of function optimization shows that with the presented methods,the searching precision is enhanced,the phenomenon of immature convergence is effectively overcome,and a satisfying optimization result is obtained.
Based on the combination of fitness property selection and elitist model, numerical crossover and uniform mutation, two methods have been put forward to improve the efficiency of real-coded genetic algorithms. The first method is eliminating the inefficient individuals in crossover operation and reserving the parents for participating selection. The second method is inserting chaos serials into the population when a trend of immature convergence appears. Simulation results of function optimization shows that with the presented methods,the searching precision is enhanced,the phenomenon of immature convergence is effectively overcome,and a satisfying optimization result is obtained.
引文
[1] 潘正君,康立山,陈毓屏.演化计算[M].北京:清华大学出版社,1998.3-4.
[2] 马钧水,刘贵忠,贾玉兰.改进遗传算法搜索性能的大变异操作[J].控制理论与应用,1998,15(3):404-408.
[3] 严 卫,朱兆达.一种遗传-梯度混合算法[J].信号处理,1998,14(增刊):42-45.
[4] 韩 炜,廖振鹏.一种全局优化算法:遗传-单纯形法[J].地震工程与工程震动,2001,21(2):6-12.
[5] 柯 珂,张世英.禁忌-递阶遗传算法研究[J].控制与决策,2001,16(4):480-483.
[6] SrinivasM,PatnaikL M.A daptive probabilities of crossover and mutation in genetic algorithms[J].IEEE TransSystMan andCy-bernetics,1994,24(4):656-667.
[7] 王 宁,蔚承建,盛昭瀚.基于嵌入混沌序列的遗传算法[J].系统工程理论与实践,1999,(11):1-7.
[8] 王东生,曹 磊.混沌、分形及其应用[M].合肥:中国科学技术大学出版社,1995.25-35.
[9] MichalewiczZ.GeneticAlgorithm s+DataStructures=EvolutionPrograms[M].Berlin(Germ any):Springer,1989.73-79.
[2] 马钧水,刘贵忠,贾玉兰.改进遗传算法搜索性能的大变异操作[J].控制理论与应用,1998,15(3):404-408.
[3] 严 卫,朱兆达.一种遗传-梯度混合算法[J].信号处理,1998,14(增刊):42-45.
[4] 韩 炜,廖振鹏.一种全局优化算法:遗传-单纯形法[J].地震工程与工程震动,2001,21(2):6-12.
[5] 柯 珂,张世英.禁忌-递阶遗传算法研究[J].控制与决策,2001,16(4):480-483.
[6] SrinivasM,PatnaikL M.A daptive probabilities of crossover and mutation in genetic algorithms[J].IEEE TransSystMan andCy-bernetics,1994,24(4):656-667.
[7] 王 宁,蔚承建,盛昭瀚.基于嵌入混沌序列的遗传算法[J].系统工程理论与实践,1999,(11):1-7.
[8] 王东生,曹 磊.混沌、分形及其应用[M].合肥:中国科学技术大学出版社,1995.25-35.
[9] MichalewiczZ.GeneticAlgorithm s+DataStructures=EvolutionPrograms[M].Berlin(Germ any):Springer,1989.73-79.
版权所有：© 2023 中国地质图书馆 中国地质调查局地学文献中心