Real-time Decision Making of Agents to Realize Decentralized Autonomous FMS by Anticipation

One of the new computer technology for controlling production systems is an autonomous decentralized Flexible Manufacturing System (FMS). The autonomous decentralized FMS aims at high production efficiency by giving self-control or decentralizing the plan, design and operation of FMS. This paper discusses the intelligent real-time decision making necessary for realizing an autonomous decentralized FMS with Automatic Guided Vehicles (AGVs) and Machining Centers (MCs). This research develops a real-time production control method called Intelligent Decisions by Anticipation and Simulations with Hypotheses (IDASH) based on the predictions that foresee not only current production situations but also anticipate future ones. IDASH can be seen that multi-production that keeps the target production ratio is possible even though neither AGV actions’ plans nor parts input schedules are given beforehand. Especially, it can be shown that the method will operate a FMS without influencing the production ratio even when unpredicted troubles happen, which is often seen in an actual factory.