Variable neighbourhood search algorithms applied to job-shop scheduling problems

Abstract

This research attempts to find the neighbourhood structures of the variable neighbourhood search (VNS) algorithm well fitting to job-shop scheduling problem (JSP). So as to achieve this objective, this research adapts the seven neighbourhood structures given by the travelling salesman problem’s literature to create the seven VNS algorithms for JSP. The performances of these proposed VNS algorithms are then compared to the performances of several existing algorithms, i.e., ACO, PSO and VNS algorithms taken from previously published literature. The comparison indicates that all proposed VNS algorithms perform better than the ACO and PSO algorithms with the significance level of 0.05. This paper finally reveals the neighbourhood structures performing well to JSP.