A Hybrid Algorithm for Solving the Unrelated Parallel Machine Scheduling Problem with Sequence Dependent Setup Times

This work addresses the unrelated parallel machine scheduling problem where setup times are sequence-dependent and machine-dependent. The maximum completion time of the schedule, known as makespan, is considered as an objective to minimize. Few studies has been done about this problem, which is a motivation to the development of this work. Other motivations are the fact that it is NP-Hard and it is commonly found in industrial processes, like textile manufacturing. Aiming to its resolution, an approach based on Iterated Local Search (ILS), Greedy Randomized Adaptive Search Procedure (GRASP) and Path Relinking (PR) are proposed. The construction phase of GRASP, which is used to build the initial solution, is inspired on the Adaptive Shortest Processing Time (ASPT) heuristic. PR is applied as an intensification strategy, after the local search. The perturbations of ILS consists in reinserting jobs from one machine to another. The developed algorithm explores the solution space using movements based on multiple insertions and swaps. It was tested using benchmark instances. Computational experiments showed that the results achieved by the proposed algorithm outperformed the results of the literature, both in terms of variability and quality of solutions.