Detecting Promising Areas by Evolutionary Clustering Search

摘要: A challenge in hybrid evolutionary algorithms is to define efficient strategies cover all search space, applying local only actually promising areas. This paper proposes a way of detecting areas based on clustering. In this approach, an iterative clustering works simultaneously algorithm accounting the activity (selections or updatings) and identifying which them deserves special interest. The strategy becomes more aggressive such detected by search. first application unconstrained numerical optimization developed, showing competitiveness method.