Performance modelling and evaluation of flexible manufacturing systems using a semi-Markov approach

This paper presents a semi-Markov modelling technique for exact and approximate analyses of queueing networks often encountered while designing Computer Integrated Manufacturing Systems. Using this approach, the states are observed only at the times of transitions. The mean residence time in each state is obtained using the probability distribution function of the conditional state occupancy times. Several appropriate states are lumped together in an effort to reduce the complexity of the original state description. The ‘lumping’ of states in order to reduce the dimensions of the problem and the achievements of bounds of performance measures result in a simple and efficient computational algorithm for this problem. Several performance measures are evaluated using the steady state probability distribution of the semi-Markov process.