Tuesday July 26/8:00
Computational Aspects of Markov Chains (Part 1 of 2)
Co-sponsored by SIAM Activity Group on Linear Algebra
Markov chains are used extensively as mathematical models throughout the biological, physical and social sciences as well as in business and engineering. As these models grow increasingly complex, the computation of the stationary and transient behavior of the Markov chain becomes increasingly difficult. This minisymposium focuses on computational problems of large-scale Markov chains and the speakers will discuss subjects ranging from computational aspects of the model building process, to promising new approaches for computing stationary and transient solutions. They will also discuss the issues of stability and sensitivity and fast techniques for determining approximate solutions of models that contain many millions of states. (For part 2, see MS27.)
Organizer: William J. Stewart
North Carolina State University
(For part 2, see MS27.)
- 8:00: Stochastic Petri Nets and Structured Markov Chains.
Gianfranco Balbo, Universita di Torino, Italy
- 8:30: Uniform Stability of Markov Chains.
Ilse C.F. Ipsen and Carl D. Meyer, North Carolina State University
- 9:00: Divide-and-Conquer Methods in Quasi-Birth-and-Death Processes.
Guy Latouche, Université Libre de Bruxelles, Belgium
- 9:30: Computing Stationary Distributions of Markov Chains by Quasi-Minimal Residual Iterations.
Roland W. Freund, AT&T Bell Laboratories