Performance evaluation of series-parallel systems

Performance evaluation of flow-type manufacturing systems is an important issue when the system is unreliable because of random breakdowns of its machines. The traditional measure of performance is the average throughput rate, but when there are delivery deadlines or production quotas to be met, the production capabilities during a time interval of fixed length become important as well. Systems with parallel machines have become more popular, and parallelism has the effect of reducing the variability of total throughput; however, to date little research has been done on the performance of such systems. In this paper series-parallel systems are modeled as continuous-time Markov chains. An efficient solution procedure is presented to estimate the throughput rate distribution, the moments of the distribution of cumulative throughput in a given period being also obtained. It is shown how these methodologies can be used during the system design stage.