Throughput Variability of Homogeneous Transfer Lines Subject to Operation-dependant Failures

Most of the research concerning production systems is so far limited to the assessment of the steady-state throughput despite the impact of the parameter variability on the overall performance of such systems. In this paper, we propose an analytical model to assess the variance of the throughput of homogeneous transfer lines subject to operation-dependant failures. The transfer lines subject of study are composed of a series of manufacturing machines with no intermediate buffers where failure and repair times are exponentially distributed. Based on Markov reward processes, the proposed analytical model determines the limiting variance of the sojourn time in the operational state. A numerical procedure has also been developed to assess the variance of the throughput for transfer lines composed of several machines with different reliability and maintainability characteristics. Numerical examples show that, although the steady-state throughput decreases monotonically, the throughout variance of transfer lines increases and then decreases as the number of manufacturing machines constituting it increases. Compared to past works considering time-dependent failure mode, the results show that in general transfer lines subject to operation-dependant failure mode are superior to those subject to time-dependant failure mode not only in term of the steady-state throughput but also when considering their respective throughput variability.