Analysis of Reconfiguration Strategies Based On Petri Nets Models and Optimization Techniques

The concept of flexible manufacturing system (FMS) has been initially introduced to enable the production of different types of parts at the same time. But dependable requirements and especially the necessity to continue to produce in spite of the breakdown of a plant component lead to exploit the flexibility to reconfigure the plant. Nowadays, the challenge for researchers and industrials is thus to build reconfigurable manufacturing systems (RMS). A RMS must be able to adapt its configuration in real-time depending on production objectives and available resources. In this paper, we introduce the concept of reconfiguration process that defines the different steps to determine the actions needed to put the system in a state compatible with production resumption. Our approach expresses the reconfiguration process as a Petri net reachability problem instance. Our proposition consists in reusing the initial models developed for control purposes and to adapt them to model the potential reconfiguration actions (alternatives routes, start and stop of resources, feasible operating sequences). We define then a target marking - not necessarily entirely known - corresponding to the specification of a family of desired system states. Then we search for the firing sequences that bring the system from an initial marking to the desired objective marking. This problem is solved using mathematical programming techniques, allowing to optimize criteria corresponding to the relevance of the reconfiguration