Intelligent decision-making for reactive scheduling in FMS

The paper describes research into the development of an intelligent simulation environment. The environment was used to analyse reactive scheduling scenarios in a specific FMS configuration. Using data from a real FMS, simulation models were created to study the reactive scheduling problem and this work led to the concept of capturing instantaneous FMS status data as `Snapshot´ data for analysis. Various intelligent systems were developed and tested to assess their decision making capabilities. The concepts of `history logging´ and expert system `learning´ is proposed and these ideas are implemented into the environment to provide decision making and control across a FMS schedule lifetime. This research proposes an approach for the analysis of reactive scheduling in an FMS. The approach and system that was subsequently developed was based on the principle of automated intelligent decision making via knowledge elicitation from FMS status data, together with knowledge base augmentation to facilitate a learning ability based on past experiences