An efficient discrete event model of assembly/disassembly production networks

This paper develops a discrete event simulation model of assembly/disassembly production networks with finite buffers and unreliable machines. The model is an extension of a previous continuous flow model, which simulates the system only when a machine's production rate is altered. The events causing changes in the production rates are: a machine fails or is repaired and a buffer becomes full, empty, not full, or not empty. During operation between two events the system runs deterministically. Thus, given the state of the system (machine cumulative production, buffer levels and their statistics) at the time of occurrence of some event, the corresponding state variables upon the occurrence of the next event can be updated using analysis. The proposed model does not use repair, not-full, and not-empty events. This is achieved by considering the machine downtimes as transient times in which no parts are produced and by developing more complex state equations. Numerical results show that the model combines speed and accuracy.