摘要
遗传算法、模拟退火算法、最优个体保护法在全局收敛性、种群早熟化、收敛速度慢等方面存在应用缺陷．文中提出了自适应退火遗传算法解决车间日作业计划的调度问题．该算法针对遗传算法中组成编码串的变异概率在整个搜索过程中是固定不变的,而且取值较小,促使算法的求解过程很长,且易走向局部最小值,提出自适应变异概率的概念与理论改善遗传算法的收敛速度;针对选择算子对种群多样性的影响,提出整体退火选择的方式(Boltzmann概率选择机制)选择杂交母体,以克服种群早熟化,避免过早收敛．最后结合车间日作业计划静态调度模型给出求解算法和求解实例．
Genetic Algorithm, Simulated Annealing Algorithm and Optimum Individual Protecting Algorithm origin from the order of nature, they exist some application limitations in the global astringency, population precocity and convergence rapidity. The Adaptively Annealing Genetic Algorithm （AAGA） is provided to deal with the scheduling question of workshop daily operating planning based on the above algorithms. In AAGA, the adaptive mutation probability is built to improve the convergence rapidity of genetic algorithm through adaptively changing mutation probability to shorten the entire optimizing process and to avoid the local optimization, the Boltzmann probability selection mechanism from simulated annealing algorithm is applied to select the crossover parents, which can solve the population precocity and the local convergence. At last, the AAGA based scheduling algorithm and domain model of workshop daily operating planning are discussed, the computing results are depicted and compared between AAGA and GA.
出处
《计算机学报》
EI
CSCD
北大核心
2007年第7期1164-1172,共9页
Chinese Journal of Computers
基金
本课题得到国家科技攻关计划项目基金(2005BA908810)
上海市世博科技专项基金(05dz05810)资助
关键词
自适应退火遗传算法
遗传算法
车间日作业计划
调度
生产计划
adaptively annealing genetic algorithm
genetic algorithm
workshop daily operating planning
scheduling
production planning
