Reduced Discrete-Event Simulation Models for Medium-Term Production Scheduling

Even today, medium-term production scheduling tends to be unsatisfactory in many industrial enterprises because it is usually based on escapist assumptions of fixed lead times and a mere summarising of the capacity supply and demand. These assumptions and the simplicity of the methods applied do not match the complexity of the scheduling problems to be solved. Thus the goals of the scheduling frequently fail to be achieved. Help might be at hand from simulation-based optimisation. Such optimisation, however, causes high computational costs, which will stand in the way of its practical application. This article reports a method of creating reduced simulation models of discrete production processes. The use of such reduced models instead of detailed ones lessens the computational costs considerably and hence makes the application of simulation-based optimisation for medium-term production scheduling possible. The reduction method was subjected to empirical investigation, the results of which are also presented here.