EquivaFormulation
收藏资源简介:
EquivaFormulation是一个开源数据集,由德克萨斯大学奥斯汀分校和斯坦福大学的研究人员创建,旨在评估优化公式等价性检查方法。该数据集包含等效优化公式及其之间的转换,通过向现有公式添加松弛变量或有效不等式生成。它用于评估EquivaMap框架的性能,该框架利用大型语言模型自动发现决策变量之间的映射,实现可扩展且可靠的等价性验证。
EquivaFormulation is an open-source dataset created by researchers from The University of Texas at Austin and Stanford University, aiming to evaluate methods for optimized formula equivalence checking. This dataset contains equivalent optimized formulas and the transformations between them, which are generated by adding slack variables or valid inequalities to existing formulas. It is used to evaluate the performance of the EquivaMap framework, which leverages large language models (LLMs) to automatically discover mappings between decision variables, enabling scalable and reliable equivalence verification.

- 1EquivaMap: Leveraging LLMs for Automatic Equivalence Checking of Optimization Formulations德克萨斯大学奥斯汀分校, 斯坦福大学 · 2025年



