A Family Competition Genetic Algorithm for Nonlinear Constrained Optimization

Abstract

Real world continuous numerical optimization problems are often heavily constrained conditions that divide into linear constraints and non-linear constraints；however, there exists no unique deterministic approach to monlinear optimization problems. This paper proposes a technique, called family competition genetic algorithm (FCGA), which incorporates the ideas of evolution strategies (ESs) and family competition into GA. The FCGA is similar to the rule of Olympic Games. The family competition con be viewed as local competition of each area. And selection is global competition in the universal tournament. The family competition can reduce the time complexity and maintain the diversity of population. Eight nonlinear constrained optimization problems are taken as benchmark problems. The results indicate that the FCGA approach is a very powerful optimization techni