AN EFFICIENT 2-STAGE ITERATIVE METHOD FOR THE STEADY-STATE ANALYSIS OF MARKOV REGENERATIVE STOCHASTIC PETRI-NET MODELS

To enhance the modeling power of stochastic Petri nets (SPNs), new ste ady-state analysis methods have been proposed for nets that include no n-exponential transitions. The underlying stochastic process is a Mark ov regenerative process (MRP) when at most one non-exponential transit ion is enabled in each marking. Time-efficient algorithms for construc ting and solving the MRP have been developed. However, the space requi red to solve such models is often extremely large. This largeness is d ue to the large number of transitions in the MRP. Traditional analysis methods require that all these transitions be stored in primary memor y for efficient computation. If the size of available memory is smalle r than that needed to store these transitions, a time-efficient comput ation is impossible using these methods. To use this class of SPNs to model realistic systems, the space complexity of MRP analysis algorith ms must be reduced. In this paper, we propose a new steady-state analy sis method that is both time and space efficient. The new method takes advantage of the structure of the underlying process to reduce both c omputation time and required memory. The performance of the proposed m ethod is compared to existing methods using several SPN examples.