Petri Nets Identification Techniques for Automated Modelling of Discrete Event Processes

Introduction

Building Petri net models from system behaviour observations is a hard task when the system is large and complex; then, the use of computer-aided modelling tools is useful. Identification techniques have been useful for building systematically models involving events and states. Finite automata and Petri nets have been used as a formalism to describe the functioning of discrete event processes in operation.

Reactive systems are embedded within an environment interacting with other systems. We focus on systems that interact through binary signals, which is the case of discrete event processes. The behaviour of the system is then captured as sequences of vectors whose entries are the values of input-output signals; afterwards, the sequences are processed by an identification method to obtain the discrete event model. This is shown in Figure 1.

Figure 1.

­

This chapter surveys relevant identification methods and overviews two approaches that generate models of different levels of abstraction; one that describes in detail the relationship between input events and outputs, and other that yields compact descriptions. Finally, current research problems and trends on discrete event process identification are discussed.