An Improved Genetic Algorithm Based on Fixed Point Theory for Function Optimization

This paper introduces triangulation theory into genetic algorithm and with which, the optimization problem will be translated into a fixed point problem. An improved genetic algorithm is proposed by virtue of the concept of relative coordinates genetic coding, designs corresponding crossover and mutation operator. Through genetic algorithms to overcome the triangulation of the shortcomings of human grade, it can start from any point to find the most advantages. Gradually fine mesh will be introduced the idea of genetic algorithms so that the search area gradually decreased, improving the efficiency of search. Finally, examples demonstrate the effectiveness of this method.