Achieving Balance Between Convergence and Diversity in Evolutionary Multi-Objective Optimization - Ke Li
Convergence and diversity are two cornerstones of evolutionary multi-objective optimization. The former means to approach the Pareto-optimal front as closely as possible, while the latter means to make the spread of Pareto-optimal solutions as uniform as possible. A good balance between convergence and diversity can ultimately provide decision makers more useful alternatives. The balance between convergence and diversity can be achieved either through the selection procedure or reproduction operator. In this talk, Ke Li introduces his recent work on how to achieve this balance by designing a selection mechanism based on stable matching and an adaptive reproduction operator based on multi-armed bandits.
In this talk, Ke Li introduces his recent work on how to achieve this balance between convergence and diversity in multi-objective optimization, by designing a selection mechanism based on stable matching and an adaptive reproduction operator based on multi-armed bandits.