Optimization of Manufacturing Cells Using Discrete Event Models

Licentiate thesis, 2006

As the name suggests, this thesis is concerned with flexible manufacturing systems (FMS) and their way of living. More specifically, the main objective of this work is to generate working schedules for the moving objects of the manufacturing cells. These schedules should be optimal in some sense, while situations involving several moving objects blocking each other or colliding should be avoided. If such working schedules could be generated automatically and efficiently, the flexibility together with the productivity of the manufacturing cells would be affected positively. To guarantee safe FMS functionment, production cells are modeled in terms of discrete event systems (DES), whereafter supervisory control theory helps to prevent undesirable behaviour. The modeling procedure is done automatically, leaving the field open for the efficiency considerations. In this work, several methods for exploring DES in search for a time optimal working sequence, based on the ideas of mixed integer linear programming, A* search and visibility graphs, are presented and studied with respect to their efficiency and robustness. In the final part of this work, the focus is shifted towards the reduction of the acceleration load imposed on the moving objects in a production cell. This leads to a decrease of strain on the manufacturing equipment, thus prolonging the productive life of the FMS. Naturally, the acceleration reduction should be, and thus is, done without compromising such qualities of the working schedule as time optimality, collision- and deadlock avoidance.