SAN lite-solver: a user-friendly software tool to solve SAN models
Afonso Sales · 2012
Structured Markovian models are widely used to map and analyze the behavior of complex systems. However, the modeler must frequently need a deeply knowledge about specialized tools or limitations imposed on the solution of their models. This paper presents an easy and practical software tool, called SAN LITE-SOLVER, that applies the Power method to solve a Stochastic Automata Network (SAN) model, using a standard Multi-valued Decision Diagram (MDD) structure to compute and to store the model’s reachable state space (RSS) and a Harwell-Boeing format (HBF) matrix which represents the underlying Markov chain (MC). The performance analysis of this new tool (in terms of memory used and CPU time to solve models) is presented and compared to the current approach used to solve SAN models. 1.