.png){kind=icon}
.png){kind=icon}
Return
Adaptive population structure learning in evolutionary multi-objective optimization
.png){kind=icon}
PRE
AI
DOI：10.1007/s00500-019-04518-x.png)
.png){kind=icon}
Original request for help.png){kind=icon}
Share.png){kind=icon}
Save
Abstract
En 中文
Some recent research shows that in multi-objective evolutionary algorithms (MOEAs), mating with similar individuals can improve the quality of new solutions and accelerate the convergence of algorithms. Based on the above finding, some clustering-based mating restriction strategies are proposed. However, those clustering algorithms are not suitable for the population with non-convex structures. Therefore, it may fail to detect population structure in different evolutionary stages. To solve this problem, we propose a normalized hypervolume-based mating transformation strategy (NMTS). In NMTS, the population structure is detected by K-nearest-neighbor graph and spectral clustering before and after the mating transformation condition, respectively. And the parent solutions are chosen according to the founded population structure. The proposed algorithm has been applied to a number of test instances with complex Pareto optimal solution sets or Pareto fronts, and compared with some state-of-the-art MOEAs. The results have demonstrated its advantages over other algorithms.
Keywords:
Evolutionary algorithm
Multi-objective optimization
Mating restriction
Population structure
AI Summary
Key information extracted from the uploaded paper, including a brief overview, abstract, background, key highlights, visual analysis, and future outlook.
Journal
IF:
2.5
Papers:
1.0W
Citations:
2.1W