Abstract: Production scheduling,job loading combination and traffic congestion are important subjective and objective factors affecting the production and operation efficiency of prefabricated manufacturing enterprises.Therefore,how to integrate production scheduling and job loading combination scheme effectively is an urgent problem for prefabricated components manufacturers with considering the time-varying characteristics of traffic congestion.To deal with the integrated optimization problem,a mathematical model for the integrated optimization problem was formulated to minimize the sum of product inventory cost,tardiness penalty and vehicle transportation cost.To reduce the difficulty of solving the problem,an Adaptive Multiple Neighborhoods Collaborative Search-based Fruit Fly Optimization Algorithm (AMNCS-FOA) was proposed.Its main characters were as follows:an integrated decision coding method with zero insertion operation was designed to represent the combination scheme of production scheduling and job loading;in the olfactory search phase of the algorithm,an adaptive probabilistic neighborhood selection strategy was proposed based on four neighborhood structures including intra group exchange,inter group exchange,intra group insertion and inter group insertion;in the visual search stage,to improve the global search ability of the algorithm,the iterative search was performed with a certain probability to accept the inferior solution as the population center.The results showed that AMNCS-FOA algorithm had faster convergence speed and better solution quality for dealing with the integrated optimization problem.Compared with the rule-based heuristics commonly used in prefabricated component manufacturing enterprises,the proposed algorithm had an average improvement rate of no less than 13% in solving quality,which was expected to significantly increase the net profit of prefabricated component enterprises and improve customer satisfaction.

Key words: precast production scheduling, time-varying transportation time, job loading combination, earliness/tardiness penalties, fruit fly optimization algorithm

摘要： 生产调度、装车组合与交通拥塞是影响预制构件制造企业生产运营效率及成本的重要主客观因素,如何在考虑交通拥塞时变特性的情况下有效集成生产调度与工件装车组合方案是预制构件制造企业迫切需要解决的问题。针对该集成优化问题,首先以最小化总提前拖期惩罚和车辆运输费用之和为目标,建立了预制构件生产调度与工件装车组合集成优化数学模型。随后为降低问题求解困难,通过深入分析问题解结构特点,提出了一种基于自适应多邻域协同搜索的果蝇优化算法(AMNCS-FOA)。其主要特点是:设计了一种带有插零操作的集成决策编码方式用于表示生产调度和工件装车组合方案;在算法的嗅觉搜索阶段,基于组内交换、组间交换、组内插入和组间插入四种邻域构造,提出了一种自适应概率邻域选择策略;在视觉搜索阶段,为提高算法全局搜索能力,以一定概率接受劣解作为种群中心进一步执行迭代搜索。计算结果显示,AMNCS-FOA算法在求解该集成优化问题时具有更快的收敛速度以及更好的求解质量。与预制构件制造企业常用的规则启发式方法相比,提出算法在求解质量上具有不低于13%的平均改进率,有望显著增加预制构件企业净利润并提高客户满意度。

关键词: 预制构件生产调度, 时变运输时间, 工件装车组合, 提前拖期惩罚, 果蝇优化算法