← All Publications

A knowledge guided transfer strategy for evolutionary dynamic multiobjective optimization

Yinan Guo and Guoyu Chen and Min Jiang and Dunwei Gong and Jing Liang

IEEE TEVC 2022 · published

Links pending

Contribution

Makes transfer decisions knowledge-guided: what to transfer, when, and how much.

Abstract

Dynamic multiobjective optimization problems exhibit patterns: certain decision regions, certain objective regions, and certain solution archetypes recur across consecutive environments. This paper formalizes these recurrences as prior knowledge and designs a knowledge-guided transfer strategy that encodes what to transfer, when to transfer, and how much to transfer based on observed similarity between historical and current environments. The resulting algorithm adapts its transfer behavior to the problem rather than applying a fixed scheme, consistently outperforming state-of-the-art baselines on synthetic and real-world DMOPs.

Why This Paper Matters

A principled way to make transfer decisions adaptive rather than fixed — a framework now widely followed in evolutionary multi-task and dynamic optimization research.

BibTeX

@article{guo2022knowledge,
  title={A knowledge guided transfer strategy for evolutionary dynamic multiobjective optimization},
  author={Guo, Yinan and Chen, Guoyu and Jiang, Min and Gong, Dunwei and Liang, Jing},
  journal={IEEE Transactions on Evolutionary Computation},
  volume={27},
  number={6},
  pages={1750--1764},
  year={2022},
  publisher={IEEE},
}