Distributed real-time expert control system

A novel distributed multiprocessing architecture for a real-time expert-based process control system is considered. A knowledge-distributed model for dynamic process control is developed. A dynamic abstraction algorithm to configure online knowledge bases and schedule inference activities which are distributed over the subexperts is described. Each subexpert focuses on a specific problem domain and operates in parallel. A novel knowledge sharing structure based on content addressable memory (CAM) to implement a novel concept called knowledge sharing with inference is explored to address time-critical intelligent control applications. An online learning and fault-tolerant structure is given for the system robustness.<<ETX>>