No label defined (Q4950474)

This is a book written for readers who have never met Markov chains before. The contents are as follows: Part I. Finite Markov chains (the background); 1. Markov chains: how to start?; 2. Examples of Markov chains; 3. How linear algebra comes into play; 4. The fundamental notions in connection with Markov chains; 5. Transient states; 6. An analytical lemma; 7. Irreducible Markov chains; 8. Notes and remarks. Part II. Rapidly mixing chains; 9. Perron-Frobenius theory; 10. Rapid mixing: a first approach; 11. Conductance; 12. Stopping times and the strong Markov property; 13. Coupling methods; 14. Strong uniform times; 15. Markov chains on finite groups I (commutative groups); 16. Markov chains on finite groups II (arbitrary groups); 17. Notes and remarks. Part III. Rapidly mixing chains: applications; 18. Random generation and counting; 19. Markov random fields; 20. Potentials, Gibbs fields, and the Ising model; 21. The Metropolis sampler and simulated annealing; 22. Notes and remark. Bibliography; Index. The list of contents points out very clearly to the author's unconventional treatment of the subject. Let us make it precise that by rapid mixing the author understands the usual loss of memory of the starting state or the number of steps performed, in certain Markov chains. (In connection with rapid mixing, the absence of any reference to Dobrushin's ergodicity coefficient is to be noted.) The readers are invited to solve 170 exercises of various degrees of difficulty, mostly easy. The book is self-contained, the only prerequisites needed are elementary probability theory and linear algebra. The presentation is very clear, with emphasis on extensive motivation of relevant ideas rather than on technicalities or completeness. In the reviewer's opinion the book should be given serious consideration for use as a graduate text for students of applied mathematics.