Constraint-Bound Agnostic Bayesian Optimization: One Model for All Thresholds

2026-07-26

AI Summary

This work addresses the challenge in industrial design where constraint thresholds are difficult to predefine and often require repeated tuning to balance feasibility and performance. To overcome this, the authors propose a constraint-boundary-agnostic Bayesian optimization framework that learns a parameterized constraint model capable of mapping arbitrary constraint thresholds to their corresponding optimal solutions. For the first time, this approach enables a single model to efficiently predict high-quality solutions under any user-specified threshold. The framework further introduces an intent-guided constraint recommendation mechanism that aligns with user preferences while enhancing objective performance. Experimental results demonstrate that the method generalizes effectively to unseen thresholds on both benchmark and real-world engineering problems, significantly reducing the need for repeated optimization and yielding superior solution quality.

Abstract

Expensive constrained optimization problems in real-world industry design often involve constraint thresholds that are difficult to determine in advance. Engineers may need to adjust constraint thresholds to explore different feasibility-performance trade-offs, requiring solutions under a wide range of threshold settings. However, existing constrained Bayesian optimization methods treat each threshold configuration independently, leading to repeated optimization and failing to exploit the shared relationship among continuously varying thresholds. To address this challenge, we propose constraint-bound agnostic Bayesian optimization (CBA-BO), a learning-based framework that learns a parametric constraint model mapping thresholds to optimal solutions. Once learned, CBA-BO directly predicts solutions for arbitrary unseen threshold configurations without additional optimization, with a one-step Bayesian optimization refinement further improving solution quality. Experiments on benchmark and engineering problems demonstrate that CBA-BO learns a transferable threshold-solution mapping, enabling efficient prediction and optimization for arbitrary threshold queries. An intent-guided constraint-bound recommendation mechanism is further developed to improve objective performance while satisfying user-specified constraint preferences.

Problem

• constrained optimization
• constraint thresholds
• Bayesian optimization
• feasibility-performance trade-offs
• threshold-solution mapping

Innovation

• Constraint-Bound Agnostic Bayesian Optimization
• threshold-solution mapping
• transferable optimization
• intent-guided recommendation
• constrained Bayesian optimization