Uncertainty-Aware Dual-Ranking Strategy for Offline Data-Driven Multi-Objective Optimization

2025-11-09

AI Summary

In offline data-driven multi-objective optimization (MOO), scarce training data leads to high epistemic uncertainty in surrogate models, and conventional Gaussian assumptions often fail. Method: We propose a dual-ranking NSGA-II framework: the first ranking employs standard non-dominance, while the second incorporates uncertainty-calibrated fitness values derived from an ensemble of uncertainty quantification techniques—including quantile regression, Monte Carlo Dropout, and Bayesian neural networks—aggregated via average rank. Contribution/Results: This approach avoids complex modeling assumptions, naturally accommodates non-Gaussian error distributions, and significantly enhances robustness and generalization under small-sample regimes. Empirical evaluation on benchmark and real-world problems demonstrates superior performance over state-of-the-art offline MOO methods, matching the efficacy of conventional surrogate-based MOO under abundant data—thereby validating its effectiveness and advancement in data-constrained settings.

Technology Category

• Machine Learning: Calibration & Uncertainty Quantification
• Reasoning under Uncertainty: Stochastic Optimization
• Search and Optimization: Sampling/Simulation-based Search

Application Category

• Search and Retrieval-Augmented AI: Web learning to rank, online learning, and counterfactual learning for ranking
• Economics, Online Markets and Human Computation: Data quality aspects of human-annotated datasets
• User Modeling, Personalization and Recommendation: Fairness-aware retrieval and ranking

Abstract

Offline data-driven Multi-Objective Optimization Problems (MOPs) rely on limited data from simulations, experiments, or sensors. This scarcity leads to high epistemic uncertainty in surrogate predictions. Conventional surrogate methods such as Kriging assume Gaussian distributions, which can yield suboptimal results when the assumptions fail. To address these issues, we propose a simple yet novel dual-ranking strategy, working with a basic multi-objective evolutionary algorithm, NSGA-II, where the built-in non-dominated sorting is kept and the second rank is devised for uncertainty estimation. In the latter, we utilize the uncertainty estimates given by several surrogate models, including Quantile Regression (QR), Monte Carlo Dropout (MCD), and Bayesian Neural Networks (BNNs). Concretely, with this dual-ranking strategy, each solution's final rank is the average of its non-dominated sorting rank and a rank derived from the uncertainty-adjusted fitness function, thus reducing the risk of misguided optimization under data constraints. We evaluate our approach on benchmark and real-world MOPs, comparing it to state-of-the-art methods. The results show that our dual-ranking strategy significantly improves the performance of NSGA-II in offline settings, achieving competitive outcomes compared with traditional surrogate-based methods. This framework advances uncertainty-aware multi-objective evolutionary algorithms, offering a robust solution for data-limited, real-world applications.

Problem

• Addresses high epistemic uncertainty in surrogate predictions
• Overcomes suboptimal results from Gaussian distribution assumptions
• Improves optimization performance under data constraints

Innovation

• Dual-ranking strategy combines non-dominated sorting with uncertainty
• Utilizes multiple surrogate models for uncertainty estimation
• Averages ranks to reduce optimization risk under data constraints