aPaRT: A Fast Meta-Heuristic Algorithm using Path-Relinking and Tabu Search for Allocating Machines to Operations in FJSP Problem

Abstract

This paper proposes a multi-start local search algorithm that solves the flexible job-shop scheduling (FJSP) problem to minimize makespan.
The proposed algorithm uses a path-relinking method to generate near optimal solutions. A heuristic parameter, $\alpha$, is used to assign machines to operations.
Also, a tabu list is applied to avoid getting stuck at local optimums.
The proposed algorithm is tested on two sets of benchmark problems (BRdata and Kacem) to make a comparison with the variable neighborhood search.
The experimental results show that the proposed algorithm can produce promising solution in a shorter amount of time.