A Constrained Genetic Algorithm Based on Constraint Handling with KS-Function and Grouping Penalty

In order to overcome the limitation when using traditional genetic algorithm in solving constrained optimization problems, this paper presents a new method of constrain handling to solve the constrained optimization problems. Firstly, the method makes full use of the condensed characteristics of the KS function to transform multi-constrained optimization problem into a single constraint optimization problem. And then a group penalty method is adopted by genetic algorithm. Aggregate constraint reduces the solution scale effectively and improves the efficiency of searching for global optimization solution. The method of penalty in grouping is used to overcome the difficulty of penalty coefficient selection for general penalty function method. Several typical numerical experiments and engineering application show the performance and effectiveness of the proposed algorithm.