A blackboard based perspective of reactive scheduling

With the challenge of rapid changes in manufacturing shop-floors, operation scheduling and executional control raise increasing demand for intelligent and responsive approaches to problems of shop floor decision making, including reactive scheduling. Timely reaction to emerging shop-floor contingencies and performance deterioration trends in executed processes, both problems are addressed by the proposed supervisory control system, SUPREACT, and its reactive scheduling function. An Expert Supervisor Unit (ESU) with its intelligent interface assists during the control of the scheduled activities’ execution and the system’s interaction with both the dynamic shop floor processes and the human supervisor. For this, the blackboard system of ESU incorporates a cognitive model of operator functions, normative to supervisory behavior in dynamic discrete manufacturing environment, and a responsive blackboard architecture. Reactive scheduling as a constraint-based opportunistic problem solving process is performed by the system’s case and rule-based reasoning knowledge sources as cooperative multi-agents and also takes into account the human’s preferences learned by the system in the intelligent distributed problem solving framework for supervisory control of dynamic manufacturing processes.