.png){kind=icon}
.png){kind=icon}
Return
A knowledge-guided regional division based evolutionary algorithm for multi-modal multi-objective optimization
.png){kind=icon}
PRE
AI
DOI：10.1016/j.asoc.2024.112059.png)
.png){kind=icon}
Original request for help.png){kind=icon}
Share.png){kind=icon}
Save
Abstract
En 中文
The characteristic of multi-modal multi-objective optimization problems (MMOPs) is that multiple equivalent Pareto solution sets (PSs) in the decision space correspond to the same Pareto front (PF) in the objective space. The difficulty in solving the MMOPs lies in how to maintain the distribution in space. Many multi- modal multi-objective evolutionary algorithms (MMEAs) take convergence as the primary selection criterion, which makes it difficult for the algorithm to find all PSs in the decision space. In view of this situation, this paper proposes a partitioned knowledge-guided MMEA with multi-stage. The algorithm makes stage changes according to the proportion of evaluation consumed by the algorithm during the evolution, and adjusts the environment selection strategy as the stage changes. At the beginning of evolution, region division is carried out to prevent the solutions on each PS from interfering with each other and evolving independently. When the evaluation consumption reaches a certain proportion, it enters the middle stage. The information of the obtained solutions are used to guide the evolutionary direction of the population, and the deleted promising solutions are reclaimed. In the later stage, the steady state updating is performed to improve the distribution of population. The experimental results on four multi-modal multi-objective test suites with different features show that the proposed algorithm is more competitive than other seven excellent algorithms.
Keywords:
Evolutionary algorithm
Multi-modal multi-objective optimization
Region division
Solutions reclaim strategy
Journal
IF:
6.6
Papers:
1.4W
Citations:
4.8W
Organization
No organization information available