Decision-Theoretic Derivation of Rule Structures for Knowledge-Based Control and Its Application to Coordinated Control of Sewage Disposal Plant

Abstract

This paper presents a method for designing a set of rules for a general knowledge-based controller. For this purpose, we first introduce a representation and analysis scheme called influence diagram from a field of the decision theory, and provide a generic model for a sensor-based control. We show that a rationally-designed control rule set can be obtained by performing probabilistic reasoning and by compiling the solutions obtained from this model. Next, as a practical application, we apply the method to the design of the knowledge-based control system for a wastewater treatment (i. e., sewage disposal) plant. We attemp to explicate the expert operator's cognitive map concerning with this plant by influence diagrams and show that our method derives a set of control rules that is rationally designed for establishing the coordinate control of multiple processes. Finally, the derived control rules are applied to an actual plant for the sewage disposal and the control result is shown.