Solving scheduling problems using Petri nets and constraint logic programming

Abstract

Formalization of scheduling problems is very often mathematically and directly bounded to a solving method. It allows one to use the properties of the problem to face the complexity of its resolution. In consequence, the initial model can not be easily extended to take into account the new constraints. Thus, the efficiency is synonym of a lack of genericity. Our starting point is the modeling of scheduling problems without any optimization objective. First, we show the advantages of Petri nets in order to study scheduling problems. Next we develop a progressive approach to design Petri net based models, through the flowshop example. Then we describe the resolution of the problem after the transcription of the obtained net into the CHIP constraint logic programming language.