An execution environment for real-time model-based supervisory control and diagnostic systems

Abstract The paper presents a Lisp based programming tool and the underlying computational model, which support implementing real-time model/knowledge-based control and diagnostic systems by giving architectural and algorithmic background for both the data processing and the scheduling aspects. The temporal characterization of the tasks and the execution requirements are generalized allowing expressing the special run-time behaviour of the artificial intelligence originated processing algorithms. The execution environment provides smooth integration of numeric and symbolic data processing methods, thus expressing the relevant knowledge -according to the model formation and execution constraints - on qualitative, quantitative and heuristic level by cooperating data processing subsystems is supported. The system architecture encourages the decomposition of the control system into loosely coupled subsystems. This feature and the message passing type of communication mechanism implemented make the distributed control system implementation possible. The basic programming features of the implementation is shown.