Genetic Algorithm And Its Application In Nonlinear Programming

Currently, in the engineering fields, especially automatic control, artificial intelligence and other areas emerge ultra-large-scale nonlinear programming problems, such problems are often multi-parameter, high complexity, uncertainty, modeling difficulties. The classic traditional algorithms for such problems have some limitations, such as getting a local optimum; computing efficiency is low, even not to be able to obtain the optimal solution.Genetic algorithm is a stochastic optimization and searching algorithm using the concept of biological natural selection and the natural genetic mechanisms, with the ability of global optimization and easy to implement. As the idea of genetic algorithm is simple, with a wide range of adaptability and high robustness, it has won a number of applications, and has become a hot research field of artificial intelligence. This article main research the improvement of genetic algorithms and research that how to play the advantages of genetic algorithm to solve the nonlinear programming problem.First of all, the background of genetic algorithm, the basic ideas and basic theory are introduced in detail. Then in view of simple genetic algorithm existence defects, such as slow convergence, local search capability not strong, poor stability, premature convergence and so on, as well as some of the existing improved genetic algorithm deficiencies, proposed a new improved genetic algorithm, and applies it in the solution nonlinear programming. Finally, by numerical experiments, we can not difficult to discover that the more complex non-linear programming problem, the more remarkable global optimization performance of the new algorithm. When dealing with non-linear planning constraints, this article uses a precise penalty function method to transform it as the unconstrained problems, then using the new algorithm proposed in this paper, in which the selection of penalty factorσsumming up a certain pattern. Finally through the examples, this article algorithm, the traditional algorithm and some genetic algorithm has carried on the comparison, the experimental results show that this constraint handing program might obtain the relatively good results.