A knowledge-based system for real-time production planning and control in a a computer integrated manufacturing system

During production, there is a need for real-time control and adaptive planning to minimize the effects of random events that affect the production capacity. Human shop floor managers must intervene and use their heuristic knowledge because the decisions are too unstructured and difficult to computerize. Knowledge-based systems are, in principle, an appropriate tool. Nevertheless, this is difficult because of the knowledge acquisition problem and the lack of suitable tools. This research investigates the use of model-based knowledge for real-time planning and control. A two-level hierarchical control system generates adaptive production trajectories using real-time information. The knowledge-based system monitors the production status, reacts to unplanned events, and performs replanning. The approach is demonstrated over a simulated cell for batch manufacturing with several workstations and material handling devices.

The results show that production goals are met with low variability and the system adapts well to temporary losses in capacity. Dynamic capacity allocation and dispatching achieve high equipment utilization. An adaptive routing algorithm, using real-time information about the traffic status, achieves low transportation delays. The shell designed for the experiments allows simple integration of heuristic and procedural knowledge for control of dynamic systems.