Linkage Learning via Probabilistic Modeling in the ECGA

The goal of linkage learning, or building block identification, is the creation of a more effective genetic algorithm (GA). This paper explores the relationship between the linkage-learning problem and that of learning probabili ty distributions over multi-variate spaces. Herein, it is argued that these problems are equivalent. Using a simple but effective approach to learning distributions, and by implication linkage, this paper reveals the existence of GA-like algorithms that are potentially orders of magnitude faster and more accurate than the simple GA.