Knee Point-Based Imbalanced Transfer Learning for Dynamic Multiobjective Optimization

Zihao Zhang; Min Jiang; Gary G. Yen

Contribution

Addresses the imbalanced distribution problem in transfer learning for dynamic multiobjective optimization by leveraging knee point information to guide knowledge transfer.

Abstract

In dynamic multiobjective optimization, transfer learning can be hindered by imbalanced distributions between source and target domains. This paper proposes a knee point-based approach that identifies critical regions of the Pareto front to guide selective knowledge transfer, mitigating negative transfer effects.

Why This Paper Matters

Identifies and addresses a key practical limitation of transfer learning in dynamic optimization — the imbalanced distribution problem — which has been largely overlooked in prior work.

When You May Find This Relevant

This work may be useful for researchers working on transfer learning in dynamic environments, particularly when dealing with distribution mismatch between optimization phases.

When handling imbalanced data distributions in transfer learning for dynamic optimization
When using knee point selection in multiobjective optimization

BibTeX

@article{zhang2024knee,
  title={Knee Point-Based Imbalanced Transfer Learning for Dynamic Multiobjective Optimization},
  author={Zhang, Zihao and Jiang, Min and Yen, Gary G.},
  journal={IEEE Transactions on Cybernetics},
  year={2024},
  note={BibTeX entry pending verification}
}

Min Jiang, Zihao Zhang, Yew-Soon Ong, Gary G. Yen · IEEE Transactions on Evolutionary Computation, 2024

Proposes a transfer learning framework that leverages historical optimization knowledge to accelerate convergence in dynamic multiobjective environments.

Spatial-temporal Knowledge Transfer for Dynamic Constrained Multiobjective Optimization

Min Jiang, Zihao Zhang, Gary G. Yen · IEEE Transactions on Evolutionary Computation, 2024

Introduces a spatial-temporal knowledge transfer framework that simultaneously exploits spatial correlations among decision variables and temporal patterns across environmental changes for dynamic constrained multiobjective optimization.

Metadata

{
  "title": "Knee Point-Based Imbalanced Transfer Learning for Dynamic Multiobjective Optimization",
  "authors": "Zihao Zhang, Min Jiang, Gary G. Yen",
  "year": 2024,
  "venue": "IEEE Transactions on Cybernetics",
  "type": "journal",
  "status": "published",
  "doi": "pending",
  "topics": [
    "Dynamic Multiobjective Optimization",
    "Learning-driven Optimization"
  ],
  "keywords": [
    "dynamic multiobjective optimization",
    "knee point",
    "imbalanced transfer learning",
    "evolutionary computation"
  ]
}