Supervisory Control of a Class of Real Time DES Based on Neural Networks Algorithm

In this paper we introduce the neural networks optimization algorithm to the discrete event systems (DES) with state transition times, which can be described hy automata model, to determine lan-guage Kopt that not only is a subset of K representing closed-loop systems's behavioar in a minimally restric-tive fashion but also makes a certain optimal performance index hold. And considered the issues related to the synthesis of supervisor using R-W theory.