Premium
Nondominated sorting genetic algorithm II with merged strategies for industrial network topology optimization
Author(s) -
Wang Junyan
Publication year - 2020
Publication title -
concurrency and computation: practice and experience
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 0.309
H-Index - 67
eISSN - 1532-0634
pISSN - 1532-0626
DOI - 10.1002/cpe.5768
Subject(s) - crossover , sorting , tournament selection , selection (genetic algorithm) , operator (biology) , genetic algorithm , mathematical optimization , computer science , cluster analysis , network topology , optimization problem , topology (electrical circuits) , mathematics , algorithm , artificial intelligence , combinatorics , transcription factor , gene , operating system , biochemistry , chemistry , repressor
Summary Fast nondominated sorting genetic algorithm II (NSGA‐II) is a popular multiobjective optimization method. However, the tournament selection strategy for crossover operator suffers from the drawback of repetitively selecting the same individuals, resulting in unsatisfying performance. To alleviate this problem, this article first proposes to employ k ‐means clustering strategy to divide the candidate individuals into multiple clusters. After that, the crossover operator are redefined with three crossover individuals, where the first and second individuals are forced to selected from the same cluster and the second and third ones are from different clusters. The newly proposed crossover operator is not only able to alleviate the phenomenon above, but also able to retain the advantage of the original tournament selection strategy. The proposed method is verified on two popular test suits, including DTLZ and ZDT test suit and an industrial network topology optimization problem. Experimental results demonstrate that the proposed method exhibits excellent performance on both the two test suits and the practical network topology optimization.