针对考虑工件尺寸不同,求解目标为极小化总完工时间的批调度问题,考虑不同的编码方式,提出了基于工件序列的蚁群算法和基于批序列的蚁群算法。基于工件序列的蚁群算法算法采用传统的工件序列编码,需要启发式规则进行分批;基于批序列的蚁群算法算法利用蚁群算法构建性编码的特点,不需要启发式规则,而采取直接分批的方式编码,充分发挥蚁群算法自身的搜索能力。针对总完工时间的优化目标,基于批序列的蚁群算法算法引入批权重构建启发式信息;针对批调度特有性质,基于批序列的蚁群算法算法加入新的信息素更新变量,设置不同的信息素初始值,并采用局部优化技术等改进措施,以克服传统蚁群算法收敛速度慢,易陷入局部最优的缺点。通过对比实验验证了所提算法的有效性。
Aiming at the problem of minimizing the total completion time for a single batch processing machine with non-identical job sizes,considering different coding methods,Job Sequence based Ant Colony Algorithm(JSACA)and Batch Sequence based Ant Colony Algorithm(BSACA)were proposed.JSACA used traditional coding method based on job sequence,batching jobs with heuristic;BSACA took advantages of constructive coding method of ACA without depending on the heuristics.It batched jobs directly and played its own optimization ability.Aiming at the optimization objective of total completion time,BSACA introduced batch-weight to construct heuristics.To overcome the disadvantages of lower convergence speed and trapping in local optimum in the traditional ACA,BSACA added a new pheromone update variable,set different pheromone initial value and adopted local optimization,which were based on the characteristics of batch scheduling.The effectiveness of the proposed algorithms was verified by comparative experiment.