Font Size: a A A

Multi-objective Function Of An Immune Genetic Algorithm

Posted on:2011-03-20Degree:MasterType:Thesis
Country:ChinaCandidate:L J ZhouFull Text:PDF
GTID:2178360332457484Subject:Applied Mathematics
Abstract/Summary:PDF Full Text Request
Genetic evolution of the abstract and the organism, by simulating the mechanisms of natural selection and genetic comes from the biological evolution a kind of "generation and detection" feature of search algorithms is formed. As the natural evolution and biological phenomena, "uncertainty", the genetic algorithm is the inevitable presence of the probability algorithm of defects.First, in the process of running of the algorithm there are two main genetic operators: crossover and mutation operator, and the conditions that the probability equals to a certain value, both of the two operators iterative search random. Therefore, the crossover operator and mutation operator provide opportunities for individuals in the evolution. However the exist inevitable emergence of mature and reduction of population in this process. Second, each practical problem to be solved has its own number of obvious features of the basic knowledge information. But the crossover and mutation operator of genetic algorithm are relatively fixed, and the probability of variability is small in the process of solving the problem. The problem ignores possible role of the characteristics of information during problem solving, especially in solving complex problems, the loss of this "neglect" is very obvious. Furthermore, the traditional genetic algorithm does not have the information storage function, which makes the workload increased dramatically when one solves the same problem the next time. The immune system is immunological memory, antibody antigen recognition and maintenance of diversity and other characteristics.In this, biological immune system is introduced into the immune genetic algorithm, the initialization technology, diversity, improving technology, and has been improved. Address four key technical point: the initialization of genetic algorithm strategy, the concentration of operators in the design and the new design of the fitness evaluation function, variation of the adaptive adjustment mechanism, optimizing the storage case, update and query. Function optimization simulation results show that, the improved genetic algorithm is more effective and feasible, them the tradition genetic algorithm. Improved algorithm not only solves the problem of degradation in the present method, but also significantly improve the convergence speed.Finally on the improved genetic algorithm a simple convergence analysis shows that the improved genetic algorithm is convergent with probability 1.
Keywords/Search Tags:Genetic algorithm, Mutation operator, Immune memory, Antigen recognition
PDF Full Text Request
Related items