An Iterated Local Search for a Re-entrant Flow Shop Scheduling Problem

Abstract

This paper discusses a re-entrant permutation flow shop scheduling problem with missing operations. The two considered objective functions are makespan and total flow time. Re-entrant flows are characterized by a multiple processing of jobs on more than one machine. We propose a heuristic for solving the problem. Since there have been promising approaches in literature on other scheduling problems, we chose the iterated local search (ILS). This meta-heuristic framework combines the advantages of local search algorithm and still tries to avoid being stuck in local optima by a so called shaking step. The initial solution for the ILS is obtained by a dispatching rule. Various rules have been tested, e.g., total job processing time and total processing time of job levels. A hill climbing algorithm has been implemented as the integrated local search method of the ILS. The ILS is compared to a MIP formulation from literature. The results show, that the ILS can deliver better results.