Agent-based dynamic scheduling for flexible manufacturing systems

The need to react to fluctuations and versatility in market demand and face competitionthreats has led to the increasing trend to produce a wide variety of product types in smallbatches. The advent of advanced technology like computerized numerically controlled (CNC)machines and automatic guided vehicles has enabled the realization of flexible manufacturingsystems (FMSs). FMSs aim at bringing the efficiency of mass production to small-to-mediumsized batch production with high product diversity. This objective calls for schedulingapproaches that optimize the utilization of the applied technology and at the same time reactto the environmental dynamics flexibly. Conventional scheduling approaches fail to provide amechanism for reacting to the dynamics of FMSs in a timely and efficient manner. Approaches that cater for optimality by a thorough investigation of available schedulealternatives always fail to exhibit real-time reactivity due to the high complexity of theproblem. In this research work, an agent-based concept for the flexible and efficient FMS scheduling isproposed. The inherent complexity of the FMS scheduling problem is tackled bydecomposing it into autonomous agents. These agents are organized in a heterarchical multi-layered architecture that builds on the flexibility of FMSs. Every involved agent appliessearch heuristics to optimize its assigned task out of its local perspective. Through theinteractions among the concerned agents along the different levels of abstraction, the scheduleis optimized from the global perspective in reasonable time. Different scheduling modes aresupported to account for the different managerial decisions and the different environmentalconditions. The generated schedule is adapted to disturbing events such as machinebreakdowns based on a schedule repair method that caters for automating the reaction todisturbances efficiently in real-time. In addition, structural changes of FMSs, including theaddition of new resources, are incorporated dynamically into the proposed scheduling, whichguarantees long-term flexibility.