Cartesian Genetic Programming Algorithm And Its Application In Art Graphic Generation

As a new kind of evolutionary art, Cartesian Genetic Programming can evolve image which has a very big change. The production created by Cartesian Genetic Programming can be used directly. It is a kind of evolutionary algorithms which has commercial value and market potential. However a series of genetic algorithm about selection, crossover and mutation of traditional Cartesian Genetic Programming is too simple. In the process of evolution, it also has an inefficient global search and easily trapping in local optimal solution. In order to solve the problems, some improved strategies are proposed. The main work of this paper is as follows.(1) Interactive Cartesian Genetic Programming are proposed based on Interactive Genetic Algorithm. Considering user’s subjective evaluations which can be gotten from interaction, they are integrated into the crossover, mutation and selection of the improved genetic operations. This makes genes more fully crossover, and mutation probability can be dynamically adjusted according to the actual condition. The whole algorithm overcomes the premature convergence of interaction, and accelerates the convergence rate. Compared with the Cartesian Genetic Programming and Interactive Genetic Algorithm, the experimental results prove that the improved algorithm has a significant improvement in the image excellence and the convergence rate.(2)The creation method of Sierpinski triangle, Koch snow, fractal tree and Mandelbrot set in fractal theory is introduced. And it is combined with the interactive Cartesian Genetic Programming. Alpha Blending is used to stack fractal graphics and the graphics evolved by interactive Cartesian Genetic Programming. The final result shows that this method can generate image with a high aesthetic value, and has research value.(3)From the development of system software, user needs are analyzed., A client is developed according to the MVC pattern design. The paper also introduced the design and development of several main functions such as reading the chromosome, generate new population, individual select/deselect, evolution and save images. Finally the software generated can satisfy the general requirements of interaction with users to help users better use interactive Cartesian Genetic Programming to image evolution.