EUVIMEDEUVIMED entdecken
Crossref

Enhancing Generalization and Scalability for Multiobjective Optimization with Population Pre-training

Haokai Hong, Liang Feng, Min Jiang, Kay Chen Tan

Evolutionary Computation · 29.05.2026 · 1-36

DOI: 10.1162/evco.a.394 ↗

Kurz gefasst

Fachartikel in der Zeitschrift Evolutionary Computation, erschienen am 29.05.2026. Verfasst von Haokai Hong und 3 weiteren Autor:innen. Beteiligte Einrichtung: The Department of Data Science and Artificial Intelligence, The Hong Kong Polytechnic University, Hong Kong SAR, P.R. China haokai.hong@connect.polyu….

AbstractSchlagwörterBibliografische AngabenZitierenVerwandte Studien

Abstract (Originalsprache)

Multiobjective optimization problems (MOPs) require the simultaneous optimization of conflicting objectives. Real-world MOPs often exhibit complex characteristics, including high-dimensional decision spaces, many objectives, or computationally expensive evaluations. While population-based evolutionary computation has shown promise in addressing diverse MOPs through problem-specific adaptations, existing approaches frequently lack generalizability across distinct problem classes. Inspired by pre-training paradigms in machine learning, we propose a Population Pre-trained Model (PPM) that leverages historical optimization knowledge to solve complex MOPs within a unified framework efficiently. PPM models evolutionary patterns via population modeling, addressing two key challenges: (1) handling diverse decision spaces across problems and (2) capturing the interdependency between objective and decision spaces during evolution. To this end, we develop a population transformer architecture that embeds decision spaces of varying scales into a common latent space, enabling knowledge transfer across diverse problems. Furthermore, our architecture integrates objective-space features through objective fusion to enhance population prediction accuracy for complex MOPs. Our approach achieves robust generalization to downstream optimization tasks with up to 5,000 dimensions—five times the training scale and 200 times greater than prior work. Extensive evaluations on standardized benchmarks and out-of-training real-world applications demonstrate the consistent superiority of our method over state-of-the-art algorithms tailored to specific problem classes, improving the performance and generalization of evolutionary computation in solving MOPs.

Schlagwörter

Stichwörter der Autor:innen

Multi-objective optimizationconstrained optimizationexpensive optimizationlarge-scale optimizationmany objectivespre-trained model

Autor:innen und Institutionen

  • The Department of Data Science and Artificial Intelligence, The Hong Kong Polytechnic University, Hong Kong SAR, P.R. China haokai.hong@connect.polyu.hk.
  • The Chongqing Key Laboratory of Big Data Intelligence and Privacy Computing, The College of Computer Science, Chongqing University, Chongqing 400044, China liangf@cqu.edu.cn.
  • The Department of Artificial Intelligence, Key Laboratory of Digital Protection and Intelligent Processing of Intangible Cultural Heritage of Fujian and Taiwan, Ministry of Culture and Tourism, School of Informatics, Xi…
  • The Department of Data Science and Artificial Intelligence, The Hong Kong Polytechnic University, Hong Kong SAR, P.R. China kctan@polyu.edu.hk.

Zitieren

APA

Hong, H., Feng, L., Jiang, M., & Tan, K. C. (2026). Enhancing Generalization and Scalability for Multiobjective Optimization with Population Pre-training. Evolutionary Computation, 1-36. https://doi.org/10.1162/evco.a.394

Vancouver

Hong H, Feng L, Jiang M, Tan KC. Enhancing Generalization and Scalability for Multiobjective Optimization with Population Pre-training. Evolutionary Computation. 2026:1-36. doi:10.1162/evco.a.394

BibTeX
@article{Hong2026,
  author = {Hong, Haokai and Feng, Liang and Jiang, Min and Tan, Kay Chen},
  title = {Enhancing Generalization and Scalability for Multiobjective Optimization with Population Pre-training},
  journal = {Evolutionary Computation},
  year = {2026},
  pages = {1--36},
  doi = {10.1162/evco.a.394}
}
Kostenlos starten →