Multiple-colony ant algorithm with forward-backward scheduling approach for job-shop scheduling problem

Abstract

The job-shop scheduling problem (JSP) is one of the strongly nondeterministic polynomial-time hard (NP-hard) combinatorial optimization problems and is dif- ficult to solve optimality for large-size problems. For practical purposes, several approximation algorithms that can find good solutions in an acceptable time have been developed. Most conventional ones in practice are based on priority dispatching rules (PDRs). In recent years ant colony optimization (ACO) has been receiving attention in solving scheduling problems including the static and dynamic scheduling problems. The successful applications of the ant algorithm to solve the static problem were founded in the single-machine weighted tardiness problem (Gaené et al. [1], Gravel et al. [2]), the flow-shop scheduling problem (Shyu et al. [3], Ying and Liao [4]), the open-shop scheduling problem (Blum [5]), and the resource constraint project scheduling problem (Merkle et al. [6]). The application to JSP has proven to be quite difficult. This paper is organized as follows. In the Section 4.2, a definition of the JSP, a graph-based representation, the general concept of ACO, the memory requirement for ant and colony, the hierarchical cooperation in multiple colonies, and the backward scheduling approach are given. The descriptions and the features in the proposed ant algorithm are presented in Section 4.3. The computational results on benchmark problems are provided in Section 4.4. Finally the conclusion and recommendation for further study are presented in Section 4.5. © 2008 Springer Science+Business Media, LLC.