PROPOSAL OF IMPROVEMENT METHOD OF NSGA-II IN MULTI-OBJECTIVE OPTIMIZATION PROBLEM

DOI IR (HANDLE)

Bibliographic Information

Other Title
  • 多目的最適化問題におけるNSGA-IIの改良方法の提案

Search this article

Description

In a multi-objective optimization problem using NSGA-II, when the number of objective functions increases, it is unlikely to work a superior relation of solutions, so the search is stagnant and it is difficult to find a proper solution. In this paper, instead of performing crowding-distance sorting and preserving the individuals with the widest distance between individuals, we adopted a method of initializing the individuals belonging to the lowest rank and generating new individuals when selecting the individuals to be preserved in the next generation among the individuals classified into the lowest rank by fast non-dominated sorting. This can be expected to overcome the search that tends to be stagnant because the superior relationship does not work by the conventional method. As a result, when there are many objective functions, a solution closer to the Pareto optimal solution can be obtained compared to NSGA-II.

Journal

Keywords

Details 詳細情報について

  • CRID
    1390855965512542208
  • DOI
    10.15002/00025389
  • HANDLE
    10114/00025389
  • ISSN
    24368083
  • Text Lang
    ja
  • Article Type
    departmental bulletin paper
  • Data Source
    • JaLC
    • IRDB
  • Abstract License Flag
    Allowed

Report a problem

Back to top