一种改进的微种群遗传算法  

An Improved Micro-genetic Algorithm

作　　者：燕乐纬[1] 陈洋洋[2] 周云[1] 

机构地区：[1]广州大学土木工程学院,广东广州510006 [2]广州大学工程抗震研究中心,广东广州510006

出　　处：《中山大学学报（自然科学版）》2012年第1期50-54,共5页Acta Scientiarum Naturalium Universitatis Sunyatseni

基　　金：国家自然科学基金资助项目(11102045);广东省自然科学基金博士启动资助项目(S2011040004039);广东省高校优秀青年创新资助项目(LYM10108)

摘　　要：采用种群隔离机制、算术交叉、杰出者保留策略等对微种群遗传算法进行了改进。减少了重启动次数,增强了两次重启动之间遗传优化过程的全局和局部搜索能力,使算法在尽可能保有模式识别信息的前提下进行智能搜索;采用了实数编码,减少了编码和解码过程中的计算开销;引入了自适应随机变异算子,使之在不增加循环次数的前提下,增加了利用现有种群已经获得的遗传信息进行有效搜索的次数;引入了异种机制,有效提高了微种群遗传算法收敛于全局最优解的概率,加快了收敛速度。最后,标准测试函数的测试结果证明了这一改进的微种群遗传算法能够用远低于标准遗传算法的计算代价获得更佳的优化效果。Population isolation,arithmetic crossover and optimum reserved strategy are used to improve micro-genetic algorithm（mGA）.Reset frequency is decreased while the global and local searching capabilities of mGA between two resets are enhanced,which makes mGA searching the parameter space intelligently as the mode recognition information is preserved as much as possible.Real-code is used to decrease the computing cost in encoding and decoding.Adaptive random mutation with existing genetic information of the current groups is used to increase efficient search.Heterogeneous strategy is used to improve the probability of convergence to global optimal solution and quicken up the convergence.Finally,standard functions testing demonstrate that the improved mGA can find better optimum solutions with less computing cost than standard genetic algorithm（SGA）.

关 键 词：微种群遗传算法 异种机制 自适应非均匀变异 算数交叉 实数编码 

分 类 号：TU318[建筑科学—结构工程] O224[理学—运筹学与控制论]