Research On Strategies Related To Evolutionary Algorithms For Constrained Optimization Problems

In the field of scientific and engineering,optimization problems are ubiquitous,but such problems often have complex constraints that complicate the search process and increase the difficulty of solving the problem.In the past few decades,evolutionary algorithms have been widely used to solve optimization problems.However,it is not suitable to solve constraint optimization problems using evolutionary algorithms alone,because they cannot directly reduce the constraint violations.Therefore,for constraint optimization problems,how to design an algorithm that can effectively deal with constraints and find the optimal solution is the focus of this paper.In this paper,from the point of constraint handling technology,combined with effective evolutionary algorithms,two improved algorithms for constrained optimization problems are proposed.The first algorithm is the robustly global optimization advantage of artificial bee colony algorithm and the stably minor calculation characteristic of constraint consensus strategy are integrated into a novel hybrid heuristic algorithm.In evolutionary search,the constraint consensus strategy is quite effective to reduce the constraint violations quickly.The performance of the proposed algorithm is verified by a set of test functions.Compared with other advanced algorithms,the experimental results show that this algorithm has good performance in optimizing quality.The second is to design a hybrid three-stage differential evolution algorithm.The core of solving constraint optimization problem is how to balance constraint violations and objective functions.When the population contains only infeasible individuals,the constraint consensus strategy is adopted at the first stage.At the second stage,the population is divided into feasible group and infeasible group.The feasible individuals are uniformly distributed in the feasible region,and the infeasible individuals are treated with an analogy penalty function method.At the third stage,the population has only feasible individuals,then the population evolves directly.The algorithm uses differential evolution algorithm as the search engine.Experiments show that the algorithm is competitive to other algorithms.