A Study on the Machine Loading Problem Considering Machine-Breakdown in Flexible Manufacturing Systems

Abstract

In this paper, the machine loading problem (MLP) as an example of combinatorial problems is presented within a Flexible Manufacturing Cell (FMC) environment. In general, MLP can be modeled as a general assignment problem and hence formulated as 0-1 mixed integer programming problem which have been effectively solved by a variety of heuristics methodologies. Recently, Neural Networks have been applied to various problems with a view of optimization. Especially, Hopfield Networks to solve MLP having objective of balancing workload of machines is discussed in this paper. While heuristic methods may take a long time to obtain the solutions when the size of problems is increased, Hopfield Networks are able to find the optimal or near-optimal solutions quickly through massive and parallel computation. By such capabilities of Hopfield Networks, it is also possible to approach real-time optimization problems. The focus in this paper is on the quality of solutions of Hopfield Networks by means of comparison with other heuristic methods through simulation, considering machine-breakdown. The simulation results show that the productivity of parts is highly improved and that utilization of each machines is equalized by achieving a balanced workload on machines.