Research On Immunity Cloning Genetic Algorithm

The immune cloning genetic algorithm has been successfully applied to some fields,such as data mining,network security,abnormal detection,optimization theory and so on.Solving constrained multi-objective optimization problems,although the immune clone genetic algorithm has excellent performance,but there are also deficiencies,such as if it is not feasible,the elite solution should not be retained,algorithm can't directly learn evolutionary experience,lack of evolutionary direction guidance mechanism.In response to the above deficiencies,this paper has improved,Specific improvement measures include:(1)On the basis of the immune clone algorithm,the introduction of environmental strategies,the definition of environmental strategies Pareto dominance and environmental strategy variability,the operating mechanism of the new algorithm is as follows:combined with the individual 4 environment information to define the environmental strategy Pareto dominance,4 environmental information including violation of the constraint degree The dominance,dominance,aggregation density,and distance from the constrained boundary,set an elite population of size N,The storage environment strategy Pareto dominates the selected individuals and then performs clone operations on the elite population.After the constraint conditions are dealt with,the environmental strategy variability is defined through the learning coefficient,the forgetting coefficient,and the repair coefficient,and the environmental strategy variation is introduced to improve the ability of the algorithm to learn prior knowledge.According to numerical experiments and quantitative metrics,the comparison results show that the efficiency of the new algorithm and the quality of the solution set has been significantly improved.(2)On the basis of the immune clone evolutionary algorithm,the introduction of population classification and direction guidance strategy,the operating mechanism of the new algorithm is as follows:Through the population classification,the antibodies are divided into non-dominated and dominant solution populations to avoid non-dominated solutions and dominant solutions.Set direct comparison,population classification,is conducive to retaining some excellent dominating solutions.The directional guidance strategy is as follows:The non-dominant solution population is used to guide the direction and distribution direction,and the dominant population is used to maintain the diversity of the population.Through directional guidance strategies,individuals are actively guided to evolve in a favorable direction as far as possible,but not completely enforced,because enforcement will lead to loss of omnipresence and globality of the population.Through numerical experiments and quantitative metrics,the comparison results show that the efficiency of the new algorithm and the quality of the solution set have been significantly improved.(3)Based on the basic genetic algorithm,the evolutionary reversal operation was introduced,and a search was performed in the vicinity of the individual and the individuals preserved were not inferior to the original individuals.New crossover and mutation operators were defined,and an improved immune clone algorithm was proposed.This paper establishes a mathematical model for avoiding reconnaissance optimal route selection and uses the new algorithm proposed in this chapter to solve the model.Test the performance of the new algorithm to provide a factual basis for the actual application value of the new algorithm.