Novel model and solution method for flexible job shop scheduling problem with batch processing machines

Abstract

This study investigates a novel flexible job shop scheduling problem, where the machines have batch-processing capacity (FJSPB). This problem is widely applied in semiconductor and other high-tech manufacturing and supply chains. In this study, a multi-commodity flow (MCF) model has been proposed which can be effectively solved to optimality by commercial solvers for small-scale problems. In addition, an improved adaptive large neighborhood search (ALNS) algorithmic framework with an optimal repair and tabu-based components (ALNSIT) is proposed, which can achieve high-quality solutions for large-scale FJSPB in a reasonable time. In the ALNSIT, a perturbation strategy and an optimal repair strategy are integrated to improve the exploitation and exploration ability of the algorithm. The proposed model and algorithms are tested on numerous existing benchmark instances. Experimental results indicate the effectiveness of the proposed model and methods, and the optimal repair strategy can significantly reduce the computational burden of the ALNS algorithm. Comparison results further verify that the proposed ALNSIT can achieve better results than existing methods for solving the FJSPB, especially for large-scale instances. Furthermore, the impacts of a wide range of features including batch capacities and instance scales on the performance of the ALNSIT and scheduling results are discussed.