The Cross‐Entropy Method: A Unified Approach to Combinatorial Optimisation, Monte‐Carlo Simulation and Machine Learning

Kybernetes

ISSN: 0368-492X

Article publication date: 1 July 2005

360

Keywords

Citation

Hutton, D.M. (2005), "The Cross‐Entropy Method: A Unified Approach to Combinatorial Optimisation, Monte‐Carlo Simulation and Machine Learning", Kybernetes, Vol. 34 No. 6, pp. 903-903. https://doi.org/10.1108/03684920510595562

Publisher

:

Emerald Group Publishing Limited

Copyright © 2005, Emerald Group Publishing Limited


Most cyberneticians and systemists have interests in simulation and in particular in fast simulation. Whilst some of the material in this book might appear to be rather specialised for many researchers in this field, there can be valuable spin‐offs that make it well‐worth reading.

The cross‐entrophy method of stochastic optimization and simulation uses an interactive procedure. Each iteration can be broken down into two phases. The authors explain that in the first phase, the procedure generates a random data sample according to specified mechanism and the second phase updates the parameters of the random mechanism based on this data. The aim being to obtain a better sample in the next iteration. And so it proceeds. The first task of the authors is to convince readers of the simplicity, efficiency and usefulness of the method for the problems that involve estimation and optimisation. It is written for a general readership of those who are interested in these problems. Other studies include;.

  • machine learning algorithms;

  • rare‐event probability estimation; and

  • combinatorial and continuous multi extremal optimization.

Related articles