Genetic Algorithm For Solving Linear Inequality Constrains Optimization Problems

Nowadays,there are many global optimization problems consisting in both the social problems and the natural science problems. In many applications,it is important to find a global minimum solution as opposed to a local solution. It means to find out the best one among the local solutions. Therefore,the global optimization is a challenge problem. Genetic Algorithm(GA) is one of the methods solving the global optimization,and it has lots of problems worth discussing.In this paper , we proposed a modified GA for solving the constrained optimization problems. The difficulty of these problems is the treatment of the constraints. Because the new individuals may be out of the feasible domain after doing the crossover operator and the mutation operator,it needs different strategies to repair the new individuals. In this paper,we design a genetic algorithm with modified repair operator. In the interest of reducing the calculation and increasing the efficiency of the algorithm,we present a new viewpoint that infeasible individuals can be optimized repaired or simple repaired with a certain proportion,and we obtain a new genetic algorithm to solve the linear inequality constrains optimization problems. This paper analyzes the convergence in probability of the new algorithm,and does the numerical value test. The result indicates that the new genetic algorithm is fit to solve a kind of linear inequality constrains optimization problems.There are five chapters in this paper. The introduction is given in Chapter 1. In next chapter,a modified genetic algorithm based on modified repair operator is proposed for solving global optimization problems. In Chapters 3 and 4,we discussed the convergence in probability and numerical results of modified genetic algorithm. The conclusion is given in last chapter.