Simulation and the Monte Carlo Method

Difficult computational problems often require solutions which adapt to the problem being solved. Such sequential methods are the focus of Simulation and the Monte Carlo Method, providing an algorithmic approach to hard counting and optimization problems, the simulation of rare-event probabilities through minimum cross-entropy, sensitivity analysis, and Markov Chain Monte Carlo. This book, by two leading experts in the field, travels well-beyond the usual introduction to stochastic simulation and variance reduction to the heart of the adaptive tools required by the complex simulation and optimization problems of the next decade. I recommend the book for researchers and practitioners alike, interested in the extraordinary power and potential of modern Monte Carlo Methods for solving problems in modeling, statistics and optimization.

This is an excellent textbook for a course on stochastic simulation for senior and master students in science. It gives a comprehensive treatment of all important aspects of dynamic discrete event simulations with examples and applications in queueing and reliability models. And each chapter concludes with many problems. In this respect it is self-contained as it has a chapter on (basic principles of) probability as well. Just a minor criticism is that the book handles traditional simulation topics such as building simulation models and verification/validation rather sketchy (in chapter 3). However, there are many other topics that you quite often do not see in books on simulation, like MCMC, optimization, rare-event simulation, cross-entropy algorithms for combinatorial optimization. The authors treat the mathematical background and details before giving the simulation algorithms, which makes the book easy to use as a reference and suitable for instruction and case studies. Specifically, I enjoyed reading the last chapter on counting problems and how to solve them (approximately) by Monte Carlo simulation. There seems to be many open problems in that area and this chapter is a good starting point for initiating interesting research.

This book is supposed to be a revision of the classical book by Rubinstein 1981. As is pointed out in the Preface: "Dramatic changes have taken place in the entire field of Monte Carlo simulation [since 1981]". This edition includes a considerable amount of new, and important, material for which the authors were among principal developers. This alone makes this book a valuable addition to the recent literature on theory and applications of Monte Carlo methods. The book is written in a clear style and is a pleasure to read.

This book is a revision of the classic first edition and is authoritative and up to date, including most of the interesting new advances in Monte Carlo methods including modern techniques like perfect sampling and Hit-and-Run algorithms.