Multi-objective job shop scheduling based on multiagent evolutionary algorithm

Abstract

With the properties of multi-objective job shop problem (MOJSP) in mind, we integrate the multiagent systems and evolutionary algorithms to form a new algorithm, multiagent evolutionary algorithm for MOJSP (MAEA-MOJSP). In MAEA-MOJSP, an agent represents a candidate solution to MOJSP, and all agents live in a latticelike environment. Making use of three designed behaviors, the agents sense and interact with their neighbors. In the experiments, eight benchmark problems are used to test the performance of the algorithm proposed. The experimental results show that MAEA-MOJSP is effective. © 2010 Springer-Verlag.