Solving unidirectional flow path design problems using metaheuristics

Abstract

Metaheuristics are widely used to solve a lot of problems for production systems. The aim of this study is to solve the flow path design problem with different metaheuristics especially developed for this particular problem. Flow path design consists of the determination of the direction of each segment and the paths that will be used by vehicles in production units. The organization of these routes directly influences the performance of the system. For example, the time of transport and the number of vehicles necessary depends on this organization. That is why this problem is one of the most important issues in AGV (automated guided vehicles) system design. This work deals with flow path design problem for a conventional unidirectional network, which is a general network where segments can be taken by vehicles in only one direction. For this specific problem, efficient optimization methods based on local search (MLS, ILS), Bee Algorithm (BA) and Ant Colony Optimization (ACO) are developed in order to solve the flow path design problem to minimize the total travel distance considering both loaded and empty trips. The problem will be described, the resolution methods will be detailed and their performances proven.