.png){kind=icon}
.png){kind=icon}
Return
Multi-objective multitasking optimization based on manifold transfer learning
.png){kind=icon}
PRE
AI
DOI：10.1007/s13042-026-03300-4.png)
.png){kind=icon}
Original request for help.png){kind=icon}
Share.png){kind=icon}
Save
Abstract
En 中文
Because of the extensive cooperative relationship between different optimization problems in real world, evolutionary multitasking optimization algorithms have received increasing attention. Knowledge transfer is the core of multitasking optimization algorithm. Knowledge from source tasks can facilitate the optimization of target tasks. Due to the differences between tasks, useless or unnecessary knowledge may be transferred in the process of evolutionary, resulting in negative transfer and reducing the efficiency of optimization. To mitigate this issue, an adaptive knowledge transfer strategy is utilized to dynamically adjust the transfer probability based on historical evolutionary information, thereby suppressing negative transfer. In order to solve the above problems, this paper proposes a multi-objective multitasking optimization algorithm based on manifold transfer learning (MOMFEA-MTL). Adaptive knowledge transfer strategy is used to adjust the probability of knowledge transfer based on historical information of population evolution. The mapping matrix is constructed by the manifold transfer method, and the solution obtained by K-means solution selection strategy is mapped from the source task to the target task to accelerate the evolution process of the population. In order to verify the effectiveness of the proposed method, experiments are carried out on 9 classical multi-objective multitasking test functions, and the experimental results show that the proposed algorithm has a good competitiveness.
Keywords:
Evolutionary computations
Multitasking optimization
Knowledge transfer
Manifold learning
Journal
IF:
2.7
Papers:
3.2K
Citations:
5.6K
Organization
No organization information available