Petri Net-Based Approach to Synthesis of Intelligent Control Systems for DEDS

Abstract

In order to control a system, especially a kind of discrete event dynamic systems (DEDS) like flexible manufacturing systems (FMS), transport systems, communication systems, etc., an additional amount of knowledge (besides the system model) is needed in the process of the control synthesis. Consequently, a suitable knowledge representation scheme must be used in addition to the standard mathematical model of the system to be controlled. Such a knowledge representation is necessary in order to express additional information concerning the particulars of the control task specifications — e.g. control aims, external circumstances and influences, different limitations, human experience, etc. Sometimes, the knowledge in question may have a logical character. However, many times its character is fuzzy. Usually, such knowledge can be expressed by means of implications in the form of the well known IF-THEN rules. A set of rules creates a knowledge base (KB) which can be utilized “off line” — i.e. during the process of the control systems synthesis — or “on line” — i.e. directly in the feedback control process of the real kind of the DEDS. The former application is illustrated by Fig. 1 and the latter one by Fig. 2. The meaning of the variables used in both figures are given in Part 3.