An Immune Recognition Based Algorithm for Finding Non-Dominated Set in Multi-Objective Optimization

An immune recognition based algorithm is proposed for finding the non-dominated set in multi-objective optimization. In this paper, non-dominated solution is comprehended from a new point of view. Then extra-optimal solution and infra-extra-optimal solution are defined and proved to be non-dominated solutions. After that, flowchart of the algorithm is outlined clearly, correctness of the algorithm is proved in detail, and computational complexity of the algorithm is analyzed theoretically. Compared with the algorithms which have the lowest computational complexity up to the present, experimental results show that when there are less objectives, the efficiency of the method based on immune recognition is almost the same as the method based on arenapsilas principle, and it is much better in the condition of more objectives.