Efficient simulation for serial production lines based on aggregated event-scheduling

Performance evaluation of complex manufacturing systems is very challenging. In many cases, simulation is one of the most important approaches available. For a manufacturing system with serial production lines, e.g. general assembly in automotive industry, the synchronous dependent machines have great impact on system performances, and increase the complexity in modeling and simulation. In this paper, we focus on developing an efficient simulation approach for a serial production line with synchronous dependent machines, where the traditional simulation methods may suffer from computation intensity. Making use of the partial system decomposability, we introduce an aggregated event-scheduling simulation approach for keeping the global event list in small size. Each section of the production line are simulated locally and synchronized through the global system states and events. A novel dividing mechanism based on max-plus algebra is introduced to guarantee the equivalence of the aggregated local simulation to the original problem. With this new approach it is possible to mimic real production systems fast and accurately within a reasonable computational time frame. Experiment results are presented to demonstrate the effectiveness and efficiency of our new simulation approach.