研究不同尺寸工件单机批调度问题,将蚁群算法与模拟退火算法相结合,引入自适应状态转移概率,提出了一种自适应蚁群退火算法AACSA(adaptive ant colony simulated annealing)。该算法利用模拟退火算法实现了一种新的混合信息素更新策略,此外根据停滞次数,动态改变状态转移概率,有效地避免算法陷入停滞以及局部最优,提高算法的性能。仿真实验结果表明,AACSA与蚁群优化算法BACO、模拟退火算法SA、启发式规则BFLPT相比,算法求解的性能更好。
This paper considerd the problem of minimizing makespan with non-identical job sizes on a single batch processing machine.Presented an adaptive ant colony simulated annealing algorithm.The algorithm adopted simulated annealing policy to implement a new mixed strategy to update pheromone,and also presented an adaptive state transition probability.This adaptive state transition probability could effectively avoid search stagnation of the algorithm.The experimental results show that AACSA has better performance than BACO(batch ant colony optimization algorithm),SA(simulated annealing) and the heuristic BFLPT.