.png){kind=icon}
.png){kind=icon}
Return
A multi-task multi-objective evolutionary algorithm for constrained multi-objective optimization problems
.png){kind=icon}
PRE
AI
DOI：10.1007/s12293-025-00471-5.png)
.png){kind=icon}
Original request for help.png){kind=icon}
Share.png){kind=icon}
Save
Abstract
En 中文
Constrained multi-objective evolutionary algorithms often struggle to simultaneously achieve both objective optimization and constraint satisfaction, particularly when dealing with multi-objective optimization problems that involve complex constraints. To address these issues, this paper proposes a multi-task multi-objective evolutionary algorithm (MTMOEA) to solve constrained multi-objective optimization problems. The algorithm includes three optimization tasks: one original task and two auxiliary tasks. Among the auxiliary tasks, one is the unconstrained optimization task with all constraints removed, and the other is the optimization task with partial constraints applied. For the unconstrained task, an optimization stopping strategy is adopted to determine whether to halt further optimization, thereby avoiding unnecessary computational resource usage. For the partially constrained task, a classification-based dynamic constraint updating strategy is employed to ensure that the task focuses on constraints that have a greater impact on problem-solving. Based on the characteristics of the three tasks, a knowledge transfer strategy, which includes two different knowledge transfer methods, is proposed. The proposed MTMOEA is evaluated against ten state-of-the-art constrained multi-objective evolutionary algorithms on three benchmark sets: LIRCMOP, DOC, and DASCMOP. Experimental results validate the effectiveness of the proposed algorithm.
Keywords:
Constrained multi-objective evolutionary algorithm
Multi-task
Optimization stopping
Dynamic constraint updating
Knowledge transfer
Journal
IF:
2.3
Papers:
447
Citations:
718