摘要
针对混合流水车间调度问题的特点,设计了一种双层编码和解码的方法,在一定程度上扩大了可行解的搜索范围,为了克服传统免疫算法在初期收敛速度慢的问题,在算法的早期应用特定的变异算子,以期提高收敛速度,为避免算法陷入局部最优解,在算法后期,利用模拟退火算法的随机扰动性,引导算法跳出局部最优解,同时还引入了免疫记忆,运用精英保留策略预防最优解的丢失,最终提出了一种改进的混合免疫算法.通过仿真结果表明了该算法在解决混合流水车间问题上的可行性和有效性.
Aiming at the hybrid flow shop scheduling problem, a new immune simulation annealing algorithm is proposed on the basis of simulated annealing algorithm and immune algorithm. This algorithm uses the method of double encoding and decoding the scope of feasible solution. Specific mutation operator can improve convergence in the early period, and the simulated annealing algorithm is presented to intensify local search ability in the latter period of the algorithm. Simultaneously, the algorithm including immunity memory and elitist strategy helps to hold the best individual. The result of the computation proves that the immune simulated annealing algorithm is feasible and practicable in solving hybrid flow shop scheduling.
出处
《大连交通大学学报》
CAS
2015年第2期96-100,共5页
Journal of Dalian Jiaotong University
基金
辽宁省百千万人才资助项目(2014921062)
辽宁省教育厅优秀青年学者成长计划资助项目(LJQ2013048)
作者简介
黄明（1961-），男，教授，博士，从事计算机管理信息系统的研究E-mail：dlhm@263．net．
