LAUSR

ABSTRACT System engineering methods for the performance evaluation of manufacturing systems require the estimation of input parameters, which is either based on real data or experts’ knowledge. In both cases, the input parameters are subjected to uncertainty. However, most of the systems engineering methods, which are based on analytical models assume that machine reliability parameters such as Mean Time to Failure and Mean Time to Repairs are precisely known. In order to overcome this limitation, this paper proposes an approach for the performance evaluation of unreliable manufacturing systems that considers uncertain machine reliability estimates. The method enables to calculate the distribution of the output performance, given the distribution of the input parameters’ uncertainty. The evaluation procedure is based on the combined use of Bayesian estimation, probability density function discretization and existing decomposition-based techniques for analyzing transfer lines composed of unreliable machines and capacitated buffers. Experimental results obtained by using the method show that neglecting uncertainty in the input parameter estimates generates significant errors in the output performance measure estimation, thus making the subsequent system operation and reconfiguration decisions sub-performing. Finally, a real case study is presented to demonstrate the potential benefits of the proposed method for real industrial applications.