Title: Diversity in multi-objective optimization: a pair-potential energy perspective.
Abstract: Generating Pareto front approximations (PFAs) with good diversity is the sweet dream of many researchers in evolutionary multi-objective optimization. However, generating such an ideal PFA, when the associated manifold is highly irregular, is challenging due to the lack of a formal definition of diversity. In this talk, we will follow an eight-year path of work around diversity using pair-potential energy functions such as the Riesz s-energy. Using these functions from physics, we can define a concept of diversity of PFAs that promotes the design of efficient subset selection algorithms to improve the performance of multi-objective evolutionary algorithms.
Bio: Jesús Falcón received a Ph.D. degree in computer science from CINVESTAV, Mexico, in 2020 under the supervision of Prof. Carlos A. Coello Coello. Jesús is currently a Research Professor with the Tecnológico de Monterrey, Mexico. He is currently supervising four PhD students and two Master's students. His current research interests are multi-objective optimization, theoretical analysis of quality indicators, subset selection, and evolutionary neural architecture search.
