.png){kind=icon}
.png){kind=icon}
Return
A direction vector-guided multi-objective evolutionary algorithm for variable linkages problems
.png){kind=icon}
PRE
AI
DOI：10.1016/j.asoc.2023.110581.png)
.png){kind=icon}
Original request for help.png){kind=icon}
Share.png){kind=icon}
Save
Abstract
En 中文
Recombination is a key component for the evolutionary algorithm to provide promising offspring solutions. However, conventional recombination operators cannot generate high-quality solutions for variable linkages problems due to the particularity of the Pareto optimal set (PS). To tackle this problem, a two-stage multi-objective evolutionary algorithm based on direction vector guidance (DSMOEA) is proposed in this paper. Firstly, a portion of the population is transformed by the eigenmatrix of the covariance matrix to increase the probability of generating high-quality offspring. Then the representative solutions are selected in the transformed population to create the direction vectors. Under the guidance of the direction vectors, the population rapidly approaches PS and generates promising offspring solutions. Finally, Differential Evolution (DE) is performed for searching globally to increase the diversity of the population. The proposed algorithm is tested on three classes of variable linkages problems with 30, 50, and 100 dimensions to verify its performance. The results show that the algorithm is promising for variable linkages problems.& COPY; 2023 Elsevier B.V. All rights reserved.
Keywords:
Multi-objective evolutionary algorithm
Variable linkages problem
Direction vector
DE algorithm
Two-stage
Journal
IF:
6.6
Papers:
1.4W
Citations:
4.8W
Organization
No organization information available