Chaos Genetic Algorithm And Its Application On Pattern Recognition

Genetic algorithm is bionic algorithm on the base of Darwin's natural evolutionism and genetic variation theory to solve the complex global optimization problems. It has been a processing optimization problem that attracts more and more attention. Because for the genetic algorithm there is a wide range feasible solution space, we avoid the partial optimal solution characteristic. But for the undulation genetic algorithm in the later period of heredity, we have the iterative number of times oversized and easy to fall into the partial optimal solution, therefore the rate of accuracy of genetic algorithm is not high.Firstly, the basic concepts of genetic algorithm and chaos theory are introduced in this thesis. After analyzing the adaptive genetic algorithm presented by M.Srinivas, a modified adaptive strategy has been proposed. Secondly, genetic algorithms based on chaotic switch and chaos optimization are studied respectively. Considering the defect of premature and slower convergence speed in genetic algorithm, by taking advantage of the stochastic, ergodic and regular properties of chaotic sequence, a improved genetic algorithm based on chaotic crossover and chaotic mutation, which is called chaos genetic algorithm, is studied. And an improved chaos genetic algorithm is proposed, which solves the function optimization problem successfully.This paper solves the optimization through programming with Visual C++ by chaos genetic algorithm in the geometric primitive extraction and simulation image segmentation. Results show that these modified methods can effectively alleviate the problem of premature convergence and greatly increase the speed of global convergence. The exploration of this paper will be beneficial for the spread of GA.