摘要
综合了遗传算法和模拟退火算法的优点,提出了一种新的遗传退火混合优化策略。该算法引入模拟退火算法作为遗传算法种群的变异算子,增强和补充了遗传算法的进化能力,同时将机器学习原理引入混合算法中,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,使系统能够在很短的时间内得到最优解。针对车间调度的典型问题进行了仿真,结果证明了新算法的有效性。
Combining advantages of Genetic Algorithm (GA) with Simulated Annealing (SA)algorithm, a new genetic annealing hybrid strategy, Modified Genetic Algorithm and Simulated Annealing(MGASA), was proposed. SA was regarded as the variation operator of GA population, which improved the local search ability and evolution. At the same time, the theory of machine-learning was introduced to MGASA, and so the average fitness of chromosomes was improved, the loss of the best solution was prevented and the speed of the evolution was increased. Then, the best solution could be obtained earlier. The simulation results of classic job-shop scheduling problems indicated the effectiveness of MGASA.
出处
《计算机集成制造系统》
EI
CSCD
北大核心
2005年第6期851-854,共4页
Computer Integrated Manufacturing Systems
基金
辽宁省教育厅资助项目(2004D113)。~~
关键词
机器学习
遗传算法
模拟退火算法
混合策略
machine-learning
genetic algorithm
simulated annealing algorithm
hybrid strategy
