The Performance of Algorithm for Solving Constrained Optimization Problems

Many engineering problems can be transformed into constrained optimization problems by establishing corresponding mathematical models. However, different solution algorithms for a constraint optimization problem usually show a significant disparity in the performance. Therefore, this paper proposes a hybrid genetic algorithm for solving constrained optimization problems. By constructing a special penalty function and combining with the random direction method, the selection, crossover and mutation operators are improved, thus enhancing the performance of the algorithm. In addition, this paper theoretically verifies the global convergence of the algorithm while employing elitism strategy.