| 基于学习机制的退火并行遗传算法应用研究 |
| |
| 引用本文： | 梁旭,黄明.基于学习机制的退火并行遗传算法应用研究[J].系统工程学报,2006,21(6):663-667. |
| |
| 作者姓名： | 梁旭 黄明 |
| |
| 作者单位： | 大连交通大学软件学院,辽宁,大连,116028 |
| |
| 基金项目： | 辽宁省教育厅计划项目(2004D1132005L036)，辽宁省基金资助项目(20052156)，大连市青年科技人才基金(2006J23JH039) |
| |
| 摘 要： | 本文综合并行遗传算法(PGA)和模拟退火算法(SA)的优点,提出一种新的退火并行混合优化策略(PGASA).该算法克服了并行遗传算法局部搜索能力弱的缺点,在子种群的搜索中引入SA作为GA种群的变异算子,增强和补充了PGA的局部进化能力;同时将机器学习原理引入到混合算法中,增加了种群的平均适值,有效地避免了最优解的丢失,加快了进化速度.针对车间调度中的典型问题进行了仿真,结果证明了新算法的有效性.
|
| 关 键 词： | 并行遗传算法 机器学习 模拟退火算法 混合策略 |
| 文章编号： | 1000-5781（2006）06-0663-05 |
| 收稿时间： | 2005-02-03 |
| 修稿时间： | 2005-02-032005-08-19 |
Application research of an annealing parallel genetic algorithm based on learning |
| |
| LIANG Xu,HUANG Ming.Application research of an annealing parallel genetic algorithm based on learning[J].Journal of Systems Engineering,2006,21(6):663-667. |
| |
| Authors: | LIANG Xu HUANG Ming |
| |
| Institution: | Software Technology Institute, Dalian Jiaotong University, Dalian 116028, China |
| |
| Abstract: | Combining parallel genetic algorithm with simulated annealing algorithm,a new hybrid optimization strategy with simulated annealing and parallel genetic algorithm is proposed.The algorithm solves the problem of weak local searching ability of parallel genetic algorithm,SA is regarded as the mutation operator of GA population,then the local searching ability is improved.At the same time,the theories of machine_learning are applied to the hybrid algorithm.The average fitness of chromosomes is improved,the loss of the best solution is avoided and the speed of the evolution is increased,then the best solution can be obtained earlier.The results are compared through the optimization calculation of the new algorithm and the traditional genetic algorithm in solving classic problem of job-shop scheduling problem,and the simulation results show the effectiveness of the new algorithm. |
| |
| Keywords: | parallel genetic algorithm machine-learning simulated annealing algorithm hybrid strategy |
| 本文献已被 CNKI 维普 万方数据 等数据库收录！ |
|
