自律分散型FMSにおける記憶と忘却を繰り返す未来予測推論

抄録

This paper describes about the method to decide hypotheses priority ranking by memory, COCOM. In order to increase the reasoning efficiency of Reasoning to Anticipate the Future(RAF)that controls AGVs(Automated Guided Vehicles)in a decentralized autonomous FMS(Flexible Manufacturing Systems). COCOM includes the memory that shows the past production conditions and the AGV actions. By using the memory, COCOM changes the hypotheses by giving the first priority ranking to the hypothesis that is likely to become a true hypothesis. COCOM was applied to the decentralized autonomous FMS which was constructed in a computer. As a result, the number of changing a true hypothesis was reduced, compared with the conventional reasoning.