The Cross-Entropy Method: A Unified Approach To Combinatorial Optimization, Monte-Carlo Simulation, and Machine Learning (Information Science and Statistics)

Rating:
Summary: The CE Method
Review: Although Cross Entropy is a relatively new methodology in optimization, there has seen an "explosion" of new articles offering theoretical extensions and new applications in the last few years. Hence, this book comes just in time to review the state of the art and help "new comers" enter this field. The method is presented in a clear, easy-to-follow manner and the best part of the book, in my opinion, is the focus on several areas of application where tough problems were already solved with CE. I have recently used this book to support a novel CE application to project management and found it extremely useful. I think it should become a standard piece in the "tool-box" of both scholars and practicionairs interested in optimization.

Rating:
Summary: A great book about a fascinating method
Review: The cross entropy method (CE) is a modern technique attacking optimization and estimation problems by simulation. It has been introduced by the first author and it is elaborated thoroughly in this book. The reader will find a lucid introductory chapter into the subject followed by the core of the book consisting of a chapter where CE returns an iterative algorithm for adaptive importance sampling simulation, and a chapter where CE is transformed into a randomized algorithm for solving combinatorial optimization problems. The book concludes with several chapters with applications including detailed numerical results and some Matlab codes.

I read the book with great pleasure because it is a well written exposition of a fascinating method containing many illustrative examples and realistic applications. I think that it is appropriate for both practitioners and theorists in simulation and optimization. While reading the book I got encouraged to apply CE to several other problems because the CE basics seems so simple while the results are marvellous. I am interested specifically in rare event simulation so I focused on reading the simulation part where I found many inspiring new ideas. In fact, I applied CE to a reliability problem and obtained results far better than existing methods. The simulation chapter is the most mathematically oriented, for instance it gives a proof of convergence and it contains recent developments in simulation of rare events with heavy tails.

I can recommend this book to everyone who likes to learn new ways for solving estimation and optimization problems.

Rating:
Summary: Just read the papers and save your money
Review: The cross-entropy method is an exciting new technique for rare event simulation and stochastic optimization.
The book unfortunately is a 99% copy and paste of the public available tutorials and papers. I bought the book before looking in the internet, so do not do the same mistake by me.

Furthermore it is quite disappointing if every chapter is written in a highly redundant manner(which follows automatically if every chapter is a paper on its own).

The topic and method is great but the book doesnt add much what the papers wont tell.

Rating:
Summary: An excellent text book for practitioners and theoreticians
Review: This is a great book intended for practitioners and "light" theoreticians. It contains precise explanations that show how to use the cross entropy method efficiently for both estimation of rare events and for optimization. The code is valuable and covers a large variety of applications. The book is deductive and easy to follow, and not cluttered with too much notations.

I really liked the applications chapters - easy to follow and show what all the fuss is about. Seems like the kind of book you'd like to have around if you're actually solving optimization problems.