Improved Cellular Genetic Algorithm And Its Application

Cellular genetic algorithm is a kind of the evolutionary algorithm combined with genetic algorithm and the principle of cellular automata, to solve the problem of lack of local search ability of genetic algorithm, so that the population diversity is maintained longer lasting and to ensure the good balance between global search and local search. But it is well known that in practice, cellular genetic algorithm in global search and local optimization between need a sufficient balance, based on PSO algorithm selection pressure is roughly reflect this balance. But often choose constant pressure while maintaining balance, but the efficiency of the algorithm is not high, a large number of evolution. By changing the ratio to change the selection pressure. I am by changing the neighborhood structure to influence the ratio, thus changing the selection pressure. The basic idea is:the reduction ratio can reduce the population pressure, promote the global search, so as to maintain the diversity of the group of higher; increase ratio can enhance the algorithm selection pressure, promote local optimization, accelerate optimization, can through influencing the ratio so as to affect the selection pressure. And ratio by the neighbor structure radius and the population grid radius, this time through at different times using different ways to define the cellular automata in neighbor, produce different neighborhood structures, ultimately choose different selection pressures at different times. Cellular genetic algorithm is a meta automaton with genetic algorithm combining evolutionary algorithms, this algorithm with genetic algorithm of wide applicability and parallel scalability and but in the late of two-dimensional cellular space diffusion speed is too slow. In this paper a multi objective cellular genetic algorithm based on three-dimensional spherical cellular space, basic idea is:take the cellular space three-dimensional ball, according to the Pareto dominance relationship find population of non dominated solutions and saved to the elite set, according to the cellular automata in topology and neighbor mechanism so that the elite of Pareto non dominated solutions to spread in a population. That index analysis and numerical experiments, the new algorithm not only diversity and uniformity is good, but also has faster diffusion in the late stage.This paper does the main work as follows:1.Briefly introduces the research history and current situation of research on genetic algorithm and cellular genetic algorithm, the main contents are summarized and the significance of the research of the.2.Analysis of the structure of cellular genetic algorithm, including cellular genetic algorithm basic principle, algorithm and parameter settings as well as the advantages and disadvantages of the algorithm.3.An overview of the steps of the algorithm the basic idea, an improved cellular genetic algorithm.4.An overview of adaptive neighborhood structure and 3D cellular space, then used for cellular genetic algorithm, and gives the detailed steps of the algorithm and the algorithm flow, finally verify the superiority of programming.