Affiliation: | aTechnical University of Lisbon, Instituto Superior Técnico, Department of Mechanical Engineering, GCAR/IDMEC, 1049-001 Lisbon, Portugal bSiemens AG, Corporate Technology, Information and Communications, Neural Computation Department, 81730 Munich, Germany |
Abstract: | This paper discusses the methodologies that can be used to optimize a logistic process of a supply chain described as a scheduling problem. First, a model of the system based on a real-world example is presented. Then, a new objective function called Global Expected Lateness is proposed, in order to describe multiple optimization criteria. Finally, three different optimization methodologies are proposed: a classical dispatching rule, and two soft computing techniques, Genetic Algorithms (GA) and Ant Colony Optimization (ACO). These methodologies are compared to the dispatching policy in the real-world example. The results show that dispatching heuristics are outperformed by the GA and ACO meta-heuristics. Further, it is shown that GA and ACO provide statistically identical scheduling solutions and from the optimization performance point of view, it is equivalent to use any of the meta-heuristics. |