A Genetic Algorithm Based on a New Fitness Function for Constrained Optimization Problem

According to the characteristics of constrained optimization problem, a new approach based on a new fitness function is presented to handle constrained optimization problems. The primary features of the algorithm proposed are as follows. Inspired by the smooth function technique, a new fitness function is designed which can automatically search potential solutions. In order to make the fitness function work well, a special technique which keeps a certain number of feasible solutions is also used. In addition, new genetic operators are proposed to enhance the proposed algorithm, i.e., crossover operator and mutation operator are designed according to whether the parent solution is a feasible solution or not. Also, to accelerate the algorithm convergence speed, one dimensional search scheme is incorporated into the crossover operator. At last, the computer simulation demonstrates the effectiveness of the proposed algorithm.