基于改进蚁群算法的多目标Job-shop动态调度

实际作业车间调度中多目标的动态优化更符合生产的需求。利用多目标优化问题的Pareto解集思想构建最大完工时间最小以及总拖期时间最小的数学模型,以事件驱动作为动态调度策略实现作业车间的动态调度。采用多目标蚁群算法优化启发式算法,并对算法的转移概率及全局信息素更新进行改进,加快算法的搜索收敛速度同时避免陷入局部最优。仿真实验证明,改进后的算法能实现Pareto前沿较好的均匀性与分布性,对双目标调度以及单个目标独自调度时的甘特图对比,表明双目标优化算法能更好地平衡各个目标的解。最后对急件插入以及机器故障两种动态事件进行仿真,验证了改进蚁群算法在实际动态调度中有较好的实现。

Multi-objective Job-shop Dynamic Scheduling Based on Improved Ant Colony Algorithm

The dynamic multi-objective optimization problem has more application in the practical job shop scheduling. This paper uses the Pareto solution of the multi-objective optimization to construct the model of the maximum completion time minimum and minimum total drag time. Then the event-driven strategy is utilized to realize the dynamic scheduling. By using the multi-objective ant colony algorithm as the optimization methods and improving the transition probability and global pheromone updating, the proposed algorithm can speed up the searching and avoid falling into local optimum. Moreover, the simulation results show that the improved algorithm can achieve better Pareto front. Compared with the Gantt chart of the multi-objective and single target scheduling, the multi-objective optimization can attain better balance among various targets. Finally, the simulations on two dynamic events show that the improved ant colony algorithm can achieve better performance in actual dynamic scheduling.