Introducing the Mallows Model on Estimation of Distribution Algorithms

摘要

Estimation of Distribution Algorithms are a set of algorithms that belong to the field of Evolutionary Computation. Characterized by the use of probabilistic models to learn the (in)dependencies between the variables of the optimization problem, these algorithms have been applied to a wide set of academic and real-world optimization problems, achieving competitive results in most scenarios. However, they have not been extensively developed for permutation-based problems. In this paper we introduce a new EDA approach specifically designed to deal with permutation-based problems. In this paper, our proposal estimates a probability distribution over permutations by means of a distance-based exponential model called the Mallows model. In order to analyze the performance of the Mallows model in EDAs, we carry out some experiments over the Permutation Flowshop Scheduling Problem (PFSP), and compare the results with those obtained by two state-of-the-art EDAs for permutation-based problems.