Reserch On The Premature Convergence Of Genetic Algorithm And Its Improved Politics

Genetic algorithm has become a hot topic in academic at home and abroad. It is a kind of advanced theory based on Darwin's theory of survival of the fittest, which produces an intelligent optimization algorithm through simulating natural biological evolution. Currently it has been successfully used in many areas which there is a wide range of applications, and the results are very good because it relies on the fitness function evaluation and doesn't need special domain knowledge besides simple in design, strong robustness, etc. It is important that genetic algorithm for evolutionary computation is one of important research branch, and attracts many scholars' attention. It gradually becomes a hot issue of research in the field of artificial intelligence.As a new technology, Genetic algorithm is short of a solid theoretical foundation and is still in period of development. It has been widely used in the actual problem, however, there are still groups of early convergence (premature), Weak capacity of local search and other issues in the application process. In view of the above problems, many scholars have proposed lots of solutions or solving strategies which had some effects. The paper undertakes a study on the genetic algorithm's "premature" based on a number of academics who have studied. The main research contents and main work as follows:1 It details introduce the genetic algorithm from the basic idea, the basic structure, and realization of genetic algorithm, and it proves the theoretical basis of genetic algorithms at the three aspects of the schema definition, assumptions and implications of building blocks such as parallel theorem.2 It details introduce genetic algorithm "precocious" phenomena of related content, especially "precocious" reason and prevent strategy. What's more, explore the inhibition of genetic algorithm "premature" phenomenon of the improvement strategy and their respective advantages and disadvantages.3 In other basis of the research, the paper also improves the standard genetic algorithm, proposes a new alternative improving algorithm AERGA through adaptive elite cross thought, fusion residual random sampling Mechanism.4 Moreover, the paper improves the standard of adaptive genetic algorithm, put forward a new algorithm RDAGA by new ideas of individual of the remaining random sampling mechanisms and dynamics of the introduction.At last, based on the test of MATLAB simulation platform for the simulation of the two kinds of algorithms, proof of this algorithm is effective, from the comparison of experimental data.