In: IEEE Trans. on Software Engineering, Vol. 24, No. 11, pages 1014-1030. 1998.
Abstract: Stochastic timed Petri nets are a useful tool in performance analysis of concurrent systems such as parallel computers, communication networks, and flexible manufacturing systems. In general, performance measures of stochastic timed Petri nets are difficult to obtain for practical problems due to their sizes. In this paper, we provide a method to compute efficiently upper and lower bounds for the throughputs and mean token numbers for a large class of stochastic timed Petri nets. Our approach is based on uniformization technique and linear programming.
Keywords: linear programming, mean token number, performance bound, stochastic timed Petri net, throughput, uniformization.