A Variable Neighborhood Descent as ILS local search to the minimization of the total weighted tardiness on unrelated parallel machines and sequence dependent setup times

Abstract

This paper addresses the total weighted tardiness minimization problem on unrelated parallel machines with sequence dependent setup times and job ready times. The problem consists in scheduling a set of jobs reducing the penalty costs caused by the delays in the job due dates. This is a NP-Hard problem and has been extensively studied in recent literature. In order to solve this, an ILS-VND hybrid metaheuristic is proposed, where a local search heuristic Variable Neighborhood Descent (VND) is integrated with Iterated Local Search (ILS) metaheuristic with multiple restarts. The results is compared with two state-of-art metaheuristics proposed in the literature. The statistical analysis indicates that for the most scenarios the proposed method outperforms the references metaheuristics.