An adaptive constraint handling technique for evolutionary algorithms

In global optimization with evolutionary algorithms constraint handling presents major difficulties, especially in the case of equality constraints. Several techniques have been proposed to overcome this difficulty. In this work an adaptive constraint handling technique is studied on a set of test problems with two evolutionary algorithms. The results indicate that the proposed adaptive technique produces results with better quality in terms of objective function values and constraint violations. The comparison was assessed by performance profiles based on a new metric that considers information both on objective function value and constraints violation.