An Flexible Selection Framework for Constrained Differential Evolution

Constrained optimization problems(COPs) is an important type of optimization problems. The swarm intelligence optimization algorithms are used to solve the problem and have achieved good results. Differential evolution(DE) is one of the algorithms and many improved constrained differential evolution(CDE) algorithms are proposed. Most of these algorithms only consider the use of constraint technology in individual updating. But like unconstrained optimization, the algorithm needs to consider the superiority and inferiority of the individuals when selecting individuals to improve performance. In order to solve this problem, an flexible selection framework for CDE is proposed in this paper. The framework measures merits and demerits of the individual from two aspects:fitness and diversity. And a flexible control method is used to control tendency of these two aspects. Experimental results show that the method has achieved good results in both the original CDE and improved CDE.