Explainable Operations Research Benchmark
收藏arXiv2025-09-30 收录
下载链接：
https://github.com/Forrest-Stone/EOR
下载链接
链接失效反馈官方服务：
资源简介：
该数据集是从开源商业项目IndustryOR中开发出来的新型基准数据集，旨在评估运筹学任务中的可解释性。该数据集涵盖了30个不同领域的问题类别，每个问题都配备了10个独特的查询。此外，该数据集通过专家评审和与实际案例的比较进行了验证，每个问题都附有Python代码，并使用Gurobi优化求解器来确定最优解。规模上，该数据集包含30个问题，每个问题有10个查询，其任务是评估运筹学中的可解释性。
This novel benchmark dataset is developed from the open-source commercial project IndustryOR, aiming to evaluate the interpretability of operations research (OR) tasks. It covers 30 problem categories across distinct domains, with each problem paired with 10 unique queries. Additionally, the dataset has been validated via expert reviews and comparisons against real-world cases. Each problem is accompanied by Python code, and the Gurobi optimization solver is used to determine the optimal solutions. In terms of scale, this dataset consists of 30 problems, each with 10 queries, and its core task is to evaluate the interpretability of operations research tasks.
提供机构：
In-house development
