Research On Immune Genetic Algorithm

Genetic algorithm is a "generation and detection" iterative process of the search algorithm. The algorithm by crossover and mutation operator achieves group search and group information exchange between individuals, in order to optimize the opportunities provided by each individual, so that the groups ensure that the direction of evolution in selection mechanism of survival of the fittest. However, in the searching process of iterative algorithm, the crossover operator and mutation operator are randomly realized under the conditions of a certain probability of occurrence. Therefore, they have provided the opportunities of individual evolution for the group, and also inevitably resulted in the possibility of degradation. Each practical question requested will have its own basic and obvious characteristics information or knowledge. However, genetic algorithm's crossover and mutation operator are relatively fixed, in solving problems, less flexible and variable, namely, they ignore a supporting role in the characteristics information of the problem, especially in solving some complex problems, this "neglect" is often brought about loss of more distinctly.In this paper, based on biological immune system's antigen recognition, maintaining the diversity of antibodies and immune memory and other features, a proposed improved genetic algorithm - the immune genetic algorithm is put forward, the algorithm will introduce the thinking of biological systems immune to the genetic algorithm, namely in use of first immune knowledge it structures inspection operator. By vaccination and immune selection, it not only retains the best individual groups but also ensures the diversity of individuals, thus avoiding the premature convergence of evolutionary search and improving convergence speed. Meantime, an improved immune genetic algorithm is designed, that is, introducing immune operator to traditional genetic algorithm, and adopting timely dynamic vaccination and the shutdown criteria are given. in the process of describing algorithm, the selection strategy of immune vaccines and the constructed method of immune operator are given. Function optimization simulation results show that with the traditional genetic algorithm, the improved immune genetic algorithm is effective and feasible. Improved algorithm can not only solve the appeared the degradation of the phenomenon of the algorithm, but also the convergence rate has markedly improved.Improved immune genetic algorithm based on vaccination conducts a simple convergence analysis and comes to the convergence of the probability of 1.