Meta-Black-Box Optimization with Ensemble Surrogate Modeling for Robustness-Accuracy Trade-off within SAEA
Researchers have developed AdaE-SAEA, a novel adaptive ensemble surrogate-assisted evolutionary algorithm designed for expensive multi-objective optimization problems. This new method integrates surrogate-assisted evolutionary algorithms within a meta-black-box optimization framework, allowing for unified control over both the infill criterion and ensemble-based surrogate modeling. AdaE-SAEA specifically addresses the robustness-accuracy trade-off in surrogate modeling by employing bagging and boosting techniques, aiming to improve exploration in early stages and exploitation in later stages of optimization. Experiments show AdaE-SAEA surpasses existing state-of-the-art and meta-black-box optimization methods, with TabPFN identified as an effective base surrogate model for ensemble learning. AI
IMPACT Introduces a novel approach to optimize complex systems, potentially improving efficiency in scientific and engineering applications.