BBOWP-Bench：评估LLM解决黑箱优化词问题的基准
原标题：BBOWP-Bench: Evaluating LLMs on Black-Box Optimization Word Problems
AI 摘要
本文介绍了黑箱优化词问题（BBOWP）这一新问题设置，要求系统从自然语言描述中推断搜索空间和优化算法。作者建立了BBOWP基准套件（BBOWP-Bench），包含数据集和评估框架，每个实例结合自然语言问题描述、可执行评估环境和人工设计的基线公式。初步评估显示，当前LLM能根据评估预算选择合适算法，但在搜索空间设计上存在困难，尤其是在问题描述信息不足或搜索空间高度特定时。
正文节选
Computer Science > Computation and Language Title:BBOWP-Bench: Evaluating LLMs on Black-Box Optimization Word Problems View PDF HTML (experimental) Abstract:Formulating an optimization problem strongly affects the quality of the final solution, yet good formulations usually require substantial expertise. Recent studies have therefore examined how to automatically derive optimization problems from natural-language descriptions, but existing benchmarks focus on settings where objective