Research On Genetic Algorithm And Its Application In Function Optimization Problems

Genetic Algorithm (GA) is a high parallel, random and adaptive searching probabilistic method based on the mechanics of natural selection and genetic. In past thirty years, genetic algorithm has been applied to many areas such as function optimization, autocontrol, machine learning etc.Function optimization is a classical application area and common example of genetic algorithm, many people pay more attention to genetic algorithm in recent years. This article mainly analyzes the mechanism of genetic algorithm, resolves the cost-effectiveness problem of medical with the Probabilistic model of Genetic Algorithm, studies the application performance of genetic algorithm in function optimization, proposes some new improved algorithms, realizes the function optimization with genetic algorithm in MATLAB.Based on the schemata theory, this article describes genetic algorithm with Markov chain and makes mathematical analysis to its convergence, lucubrates the influences of constitutive factors to genetic algorithm in applications.According to the achievement of Probabilistic model genetic algorithm, this article studies the algorithm principium of different types of PMBGA and respective characteristic, applies the Bayesian model to the analysis of complex medical cost-effectiveness and makes well performance.To overcome the empiricism and blindness of parameters selection when people resolve problems with genetic algorithm, this article presents a method to search the best combination of crossover and mutation ratio to resolve function optimization with dynamic parameters, intuitionistic relation graphic of parameters and statistical analysis. At the same time, this article also proposes a high-performance adaptive GA which adjusts the probability of crossover and mutation ratio according to the population fitness, the algorithm is proved effectively by experiments.By the MATLAB optimization toolbox, one-dimension and multi-dimension function optimization is solved effective with GA, straight graphics, data and better performance can also be obtained.