ABSTRACT

This article presents a flexible job-shop scheduling problem with a parallel assembly stage and lot streaming. Suppose that several products of different kinds are ordered to be produced. Each product consists of several specific parts. The components (parts) of products are manufactured in a flexible job shop, then assembled into products in parallel lines or parallel stations. The objective function is to minimize the total completion time of products (makespan). First, the problem is described and the parameters and decision variables are defined. Then, the problem is modelled as a mixed-integer linear programming model to solve the problem using GAMS software. Owing to the NP-hardness of the problem, two new algorithms are proposed to solve it for medium- and large-sized instances. These algorithms are based on variable neighbourhood search (VNS), with self-adaptive parallel VNS being applied in one of them. The considered problem is decomposed into two subproblems and each algorithm is used on two levels. These algorithms are applied to solve test problems of different sizes. The results show that the two-level algorithms perform better than the integrated one and the algorithm with self-adaptive parallel VNS outperforms the other algorithms in terms of solution quality.

摘要

本文介绍了一个具有并行装配阶段和批次流的灵活作业车间调度问题。假设订购了几种不同种类的产品。每个产品都由几个特定的​​部分组成。产品的组件（零件）在灵活的工作车间中制造，然后在平行线或平行工位中组装成产品。目标函数是最小化产品的总完成时间（makespan）。首先，描述问题并定义参数和决策变量。然后，将问题建模为混合整数线性规划模型，使用 GAMS 软件解决问题。由于问题的 NP 难度，提出了两种新算法来解决中型和大型实例的问题。这些算法基于可变邻域搜索 (VNS)，其中一种算法应用了自适应并行 VNS。所考虑的问题被分解为两个子问题，每个算法在两个层次上使用。这些算法用于解决不同规模的测试问题。结果表明，两级算法的性能优于集成算法，具有自适应并行VNS的算法在解决方案质量方面优于其他算法。