Optimising the JIT methodology using distributed artificial intelligence techniques

The authors discuss the way in which distributed AI techniques can handle both the one-off problem of planning the layout of a factory appropriate to just-in-time operation, and the problem of scheduling and controlling just-in-time manufacturing at the process level and in real time. It is demonstrated that in order to obtain the necessary accuracy and speed of response, the AI-based system must be distributed and localized. This concept is called the localization principle. A two layer model of localized distributed AI and an appropriate system architecture are presented