To respond rapidly to the highly volatile market, the reconfigurable manufacturing systems (RMS) have brought forward challenging issues. First of all there is a need to build a formal model of a manufacturing configuration. Second it has to be rather easy to derive the models associated to the manufacturing configuration changes (reconfiguration) from such an initial model. An off-line method of rapid design of an optimal logic control law (configuration) based on Petri net (PN) is presented in this paper. From a controlled system modeling point of view, the main characteristics of the level 1 of the CIM architecture are depicted. Subsequently, the formal tool used in the automated planning field is extended to provide a controlled system model. The concept of operation is structured in order to introduce the behavioral properties of the operations. A four-step method is then proposed to design a logic control law that satisfies several goals: reduction of the lead time, satisfaction of the work orders objectives, minimization of the time cycle. Finally, the proposed design method is illustrated on a manufacturing cell.