In: 1st Int. Workshop on Manufacturing and Petri Nets, 17th Int. Conf. on Appl. and Theory of Petri nets, pages 215-234. June 1996. Presented but not published; available at http://www.is.titech.ac.jp/labs/yukiolab/doctor/marco/research.html.
Abstract: We present SNS, a Solver for Synchronized Networks modelled as generalized stochastic Petri nets. General enabling and firing functions and reward measures are specified in the C programming language, providing extensive flexibility with respect to conventional net descriptions.
Unstructured and structured steady-state and transient solutions are possible, to compute the specified rate-based and impulse-based reward measures. We allow the model to have both immediate and timed synchronizing transitions. Furthermore, the user can solve numerical linear algebra problems by a variety of algorithms based on solution vectors and matrices. We apply SNS on a case study of kanban systems and give numerical results. In the conclusion, we point out ideas to expand the package.
Keywords: generalized stochastic Petri nets, numerical solution of (structured) Markov chains, Markov reward models, Kronecker algebra, numerical linear algebra.