The analysis of an unreliable two-machine production line with random processing times

This paper develops an efficient algorithm for the analysis of an unreliable two-machine production line with random processing times and finite buffer storage between the machines. The two-machine line is modelled by a continuous time Markov process and an efficient analytical method is developed to derive the steady state distributions of the line. With these steady state distributions, the steady state output and the average inventory level can be easily determined. Numerical experiments show that the algorithm is simple, accurate, fast, and requires less computer memory compared with the currently available methods.