基于离散粒子群算法的炼钢连铸生产调度

Discrete Particle Swarm Optimization for Steelmaking and Continuous Casting Production

在炼钢连铸生产调度中,最大完工时间优化可保证生产计划准时完成,总完工时间优化可降低板坯存储过程中的物耗和能耗。结合生产实际,提出了以最小化总完工时间为目标,最大完工时间为目标约束的处理方法。首先根据改进NEH算法确定最大完工时间的可行上界,再利用离散粒子群算法对总完工时间进行优化。在算法设计过程中,通过加入启发式规则提高初始解性能,使用交换操作实现位置更新公式的离散化,引入Metropolis准则避免算法陷入局部最优。最后通过案例分析,验证了求解方法的可行性和有效性。

In the scheduling of steelmaking and continuous casting production, it can ensure timely completion of the production plan and reduce material consumption and energy consumption by optimizing makespan and total completion time respectively. Considering the actual situation of the workshop, a method which using makespan as a constraint when scheduling total completion time is proposed for the problent Firstly, a feasible upper bound of makespan is discussed based on NEH algorithm, then a discrete particle swarm optimization （DPSO） /s nsed to optimize the total completion time. The DPSO algorithm is designed, in which the NEH rule/s integrated in the initialization process and several composite exchange operations are exploited for location update formulca Furthermore, the Metropolis criterion is adopted？to avoid plunging local optimum Finally, experimental comparisons demonstrate efficiency and competence of the method.