NextEval improves optimization of costly black-box problems with AI agent
NextEval: An Agent-Based Solver and Benchmark for Derivative-Free Optimization
Machine LearningArtificial Intelligence
Summary
Sometimes, people need to find the best settings for a problem but can't use common tricks because they can't easily measure how things change. The authors created NextEval, which uses an AI language model to decide where to test next in order to learn the most with a limited number of tries. This AI works together with traditional math methods, sharing information to pick better test points. Their tests show NextEval solves these problems more efficiently than previous best methods.
What this means in practice
- •For simulation engineers: Optimize simulation parameters more efficiently when derivative information is unavailable or expensive to acquire.
- •For experimental design teams: Select the most informative experimental trials within a limited testing budget by guiding trial selection with AI.
Authors
Shuo Chen, Li Cui, Cunxin Huang, Ya-Feng Liu, Minghang Xie, Meng Xu, Zaikun Zhang
Abstract
Derivative-free optimization (DFO) addresses problems where derivatives are unavailable or costly to obtain. This setting often arises in simulation-based and experimental applications, where evaluating the black-box objective can itself be costly. Thus, it remains a central challenge in DFO to effectively exploit accumulated observations to select informative evaluation points under a limited budget. In this light, we introduce the NextEval solver, an agent-based DFO solver, which integrates the capabilities of a large language model (LLM) to guide the sequential selection of evaluation points. Specifically, we establish a system that combines an LLM agent with a numerical environment. The environment provides tools for analyzing past evaluations and executing numerical methods, while maintaining a shared evaluation history and method-specific state to support continuation across method switches. Using observations and tool feedback, the LLM agent selects tools, configures their parameters, and allocates the evaluation budget, while numerical methods compute and evaluate candidate points. Building on this system, we also introduce NextEval Bench, a benchmark protocol for comparing agent configurations on shared DFO tasks using the same numerical tools under matched evaluation budgets and information constraints. Numerical results on low-dimensional DFO problems demonstrate that the NextEval solver outperforms the state-of-the-art DFO baselines in function-evaluation efficiency.