Efficient Representations of Alternative Models of Discrete Event Systems Based on Petri Nets

Abstract

The design of technological systems that can be modelled as discrete event systems (DES) constitute a difficult problem with important implications in the performance of the operation of the systems during their complete lifetime. The independent simulation of the behaviour of selected alternative structures for the system to be designed constitutes a classical approach for the decision support systems. This paper provides with a new methodology for affording the design process of a DES that overcomes two of the main drawbacks of the classical approach. On the one hand, simulation is substituted by optimisation, hence it is a search algorithm that chooses the operation conditions of the simulations. On the other hand, the independent analysis of a selected set of alternative systems that comply with the specifications of the system to be designed is transformed in a single process, where the independent state spaces are merged and the redundant information on the different alternative models is reduced. As a result, an important improvement in the efficiency of the algorithms developed for the decision support systems is achieved. An introduction in the concepts of alternative and compound Petri nets, as well as the transformation algorithms is provided in this paper. Furthermore, an illustration of the mentioned concepts is given. The conclusions reveal this approach as very promising, applied to the design of the structure of discrete event systems.