Optimal Problems Research Based On Artificial Immune Algorithm

An improved immune genetic algorithm is presented by research of Biological Immune Systems' (BIS) response process to exotic antigens intrusion in this paper. And this algorithm's characteristic and validity are both analyzed. Its convergence is also approved. The detail work as follows,Firstly, I mmune Genetic Algorithm (IGA) based on the selection probability of similarity and vector distance is proposed using density regulating theory, individual diversity retaining strategy and immune memory function. The algorithm can effectively o vercome immature convergence phenomenon in Simple Genetic Algorithm (SGA). It can improve not only antibody's similarity but also its diversity. And it can avoid local optimal solution and shorten searching time. At the same time this paper presents a general expressing form of this kind of the algorithms.Secondly, this paper takes the immune vaccine's idea into IGA based on selection probability of similarity and vector distance, which retains many good performances. This algorithm restrains degenerative phenomenon in the process of optimization using problems' characteristic information or prior knowledge deliberately. New way to get vaccine that retaining better antibodies' characteristic information makes algorithm convergence in the whole process rapidly.Lastly, this paper applies IGA to Traveling Salesman Problem (TSP), a kind of typical combinational optimization problems. During solving the problem, presents a different code strategy from other general ones, discusses some important factors that influence the algorithm characteristics and gets the ranges of the parameters b y experiments. C omparing the a lgorithm with o ther s ix algorithms, the r esults s how that the convergent speed of the algorithm is faster than others.