An approach for Assembly Sequence Planning based on Max-Min Ant System

This paper presents an approach for assembly sequence planning based on Max-Min Ant System. The research proposes a metaheuristic for minimizing the number of reorientation and tool changes, to decrease the total operations during an assembly process. The work considers six possible assembly operations along the positive and negatives axes. To guarantee the validity and feasibility of sequences a disassembly matrix is used in the proposal. It is also used to construct dynamically the disassembly completed graph of the problem and for saving memory in its representation. The main contribution of this paper is the resolution of the problem by the Max-Min Ant System considering four constraints: number of reorientations, tool changes, precedence and geometric constraints. The approach performance is verified through case studies, most reported in literature. As fundamental conclusion we drew that the tool change constraint can be addressed in the proposed metaheuristic by including it in the probabilistic rule.