@article{Li2026, 
author = {Wei Li and Weize Qin and Ziqiong Liu and Ning Yang},
title = {Constraint Ranking and Multitask-Assisted Algorithm for Constrained Multi-Objective Optimization},
year = {2026},
journal = {Complex System Modeling and Simulation},
volume = {6},
number = {4},
pages = {447-464},
keywords = {constrained multiobjective optimization, constraint handling, evolutionary algorithm, constraint ranking, evolutionary multitask optimization},
url = {https://www.sciopen.com/article/10.23919/CSMS.2025.0024},
doi = {10.23919/CSMS.2025.0024},
abstract = {Most existing studies focused on the conflict between objective optimization and constraint satisfaction when solving constrained multi-objective optimization problems (CMOPs). However, as constraints become increasingly varied, they can exert different influence levels on the optimization process. Addressing this gap, this article proposes a constraint ranking and multitask-assisted constrained multi-objective evolutionary algorithm (CR-MT-CMOEA). CR-MT-CMOEA proposes a constrained ranking strategy to rank all constraints based on their influence levels, where a constraint with a greater influence will be assigned a higher rank and handled in a higher priority. After the constraint ranking, CR-MT-CMOEA adopts a multi-task optimization framework to handle all ranked constraints one by one by based on their priorities. The multitask-assisted framework could dynamically relax the constraint boundary and enhance the population convergence and diversity through co-evolution. The constrained ranking strategy and the multi-task optimization framework ensure that CR-MT-CMOEA could effectively solve CMOPs with multiple complex constraints. CR-MT-CMOEA is tested against seven CMOEAs on three test suites and 12 real-world problems. The experimental results demonstrate that CR-MT-CMOEA has significantly better performance.}
}