Automatic Symbolic Solution of Markov Chains

摘要

Continuous time Markov chains are commonly used in system performance modeling. Increasing system complexity and non-Markovian behavior can drastically increase the size of a Markov model's state space. Accordingly, approximation techniques have been introduced to reduce the resources needed to solve Markov chain models. In this paper the authors discuss a method for automatically deriving symbolic solutions of Markov chains. Symbolic solutions should provide insight when attempting to evaluate the validity of both Markov models and approximation techniques for their solution. (Author).