Theoretical and Empirical Investigations on Difficulty in Structure Learning by Estimation of Distribution Algorithms

Estimation of distribution algorithms (EDAs) are population based evolutionary algorithms derived from genetic algorithms (GAs) . EDAs build probabilistic models of promising solutions to guide further exploration of the search space. They have been considered to behave in similar way to GAs. In this paper, we show their different behaviors and difficulties in applications of EDAs by designing an EDA difficult function in which schemata that are not consistent with problem structure sometimes overwhelm those that are.