Abstract: The distributed hybrid flow shop scheduling problem and assembly shop problem are widely exist in real manufacturing systems.In actual production,besides machine resources,worker resources are also key factors affecting production efficiency.Therefore,the energy-efficient distributed assembly hybrid flow shop scheduling problem was studied considering worker resources.A mixed integer linear programming model was established aiming at minimizing total tardiness and total energy consumption.Based on the characteristics of the problem and its multi-objective nature,a Q-Learning Memetic Algorithm (QLMA) was proposed.In QLMA,to generate excellent initial solutions,a feature-based initialization strategy was introduced.Meanwhile,a Q-learning based variable neighborhood local search was proposed to refine non-dominated solutions,thereby guiding population evolution.Furthermore,an energy-saving strategy was designed to further optimize total energy consumption.Finally,extensive experiments were conducted on 90 large-scale instances,and compared with three other algorithms,the results demonstrated the effectiveness of the proposed algorithm.

Key words: distributed hybrid flow shop scheduling problem, assembly shop problem, worker resources, energy-efficient, Q-learning memetic algorithm

摘要： 分布式混合流水车间调度问题和装配车间问题已广泛存在于现实的制造系统中。 在实际生产中,除机器资源外,工人资源也是影响生产效率的关键因素。因此,研究了考虑工人资源的分布式装配混合流水车间节能调度问题。 首先,建立以最小化总延迟和总能耗为目标的混合整数线性规划模型。 基于问题特征及多目标特性,提出了一种Q-learning模因算法 (QLMA)。 在QLMA中,为了生成优秀的初始解,提出了一种基于问题特征的初始化策略。同时,采用一种基于Q-learining的变邻域局部搜索对非支配解进行细化,从而引导种群进化。 此外,设计了一种节能策略,以进一步优化总能耗。最后,在90个大型实例上进行了大量的实验,并与其他3种先进算法进行比较,验证了QLMA算法的有效性。

关键词: 分布式混合流水车间调度问题, 装配车间问题, 工人资源, 节能, Q-learning模因算法