Font Size: a A A

Research Of The Population’s Diversity Of Artificial Immune Algorithm

Posted on:2014-02-04Degree:MasterType:Thesis
Country:ChinaCandidate:S Y WangFull Text:PDF
GTID:2248330395483836Subject:Information networks
Abstract/Summary:PDF Full Text Request
With the development of information technology, multi-objective optimization have a newsignificance recent years,and become an important researching direction in optimizing field. Withyears of development, Artificial Immune System is the computing system based on the theory ofnatural immune, the researchers designed a large number of immune algorithm to solve themulti-objective optimization problem, and showed good performance.However,the study found that all kinds of immune algorithms’ operational principles andprocess are not same. Thus, in order to improve our understanding of the meanings of immunealgorithm,and to use them more effectively, it is necessary to make a study of artificial immunesystem. This paper presents a method that can unify the expression of the multi-objectiveoptimization immune algorithm, and abstracts the main principle and operational process of threekinds core operators of the immune algorithm.Because multi-objective optimization immune algorithms’ goal is to get the right Pareto set andPareto optimal solutions,which is maintaining good closely and distribution. And, when we getgreat diversity of the population, it’ll help algorithms to find the potential optimal solutions, and toensure the solutions concentrate better distribution. So, from protect population diversityperspective, the researchers designed a large number of immune algorithms to solve themulti-objective optimization problem. It is heavily based on three immune operators: the cloneselection operator, the hyper-mutation crossover operator, the recombination and memory operator.Then, analysis of three kinds of operators how or when the impact on the diversity of the populationimmune algorithm, which can show that the immune operator have the link between the diversity ofthe population. Motivated experimental analysis of MISA, NNIA and CMOIA three classic immunealgorithms in two reference index values indicates that the clone selection operator with a negativeimpact on the algorithm diversity, but the hyper-mutation crossover operator or the recombinationand memory operator child will help the algorithm to maintain diversity.
Keywords/Search Tags:Multi-objective optimization, artificial immune system, immune operator, diversity
PDF Full Text Request
Related items