A PID-KNOWLEDGE-BASED SYSTEM FOR NONLINEAR PROCESS-CONTROL

An object-orientated knowledge-based management system (KBMS) architec ture based on the concept of proportional-integral-differential (PID) control with self-learning capabilities is proposed for the high level control of complex systems. The architecture comprises: an enhanced K BS with tight coupling to a deductive database which allows both botto m-up and breadthwise search strategies to be implemented in an optimal combination. This management system selects both the set-point and th e independently controllable variables; the independent variables can then be modified by an inductive process. The resultant system utilise s experimental data for the tasking and control of a CO2 laser manufac turing system. (C) 1995 John Wiley & Sons, Inc.