General Toolkit for Discrete Estimation of Distribution Algorithms

Abstract

This paper presents a general toolkit developed for discrete Estimation of distribution algorithms(EDAs). The essense and procedure of each algorithm is presented, and the toolkit is used to solve several toy problems, which is presented in the latter part of the paper. We implemented and presented a general toolkit for EDAs(gtEDA) as a Matlab toolkit for EDA experiments. The experiments conducted using this toolkit have demonstrated the performance of different algorithms. However, to make the toolkit more complete, the extension to multivariate dependencies should be developed, and the algorithms should be extended to continuous variants.