A fast and efficient analytical method for throughput evaluation of unreliable series-parallel production lines

Abstract

In this paper, an efficient throughput evaluation method for large series-parallel production lines is developed. The parallel machines follow the exponential reliability model and are not necessarily identical. The proposed method considers different aggregation techniques to transform each group of parallel machines into an equivalent station. Then, the throughput of the obtained buffered serial production line is evaluated using an analytical approach. This evaluation method is based on the analysis of buffer states using a birth-death Markov process. The strength of the approach lies in the analysis of only full and empty buffer states rather than all buffer states. Consequently, through the comparative numerical study against simulation and recent literature, the proposed approach demonstrates its accuracy and efficiency in evaluating throughput of large complex series-parallel systems in only few seconds. In addition, aggregation techniques used are compared and some relevant theoretical properties are derived.