DACBench

DACBench是一个标准化的基准库，旨在收集和标准化来自不同AI领域的现有DAC基准，并提供新基准的模板。该数据集由信息处理研究所(tnt)和汉诺威莱布尼茨大学联合创建，包含6个初始的多样化DAC基准，涵盖AI规划、深度学习和进化计算等领域。DACBench的设计重点在于灵活性、可重复性、可扩展性以及自动文档和可视化。数据集的应用领域广泛，旨在通过动态控制目标算法的高级参数设置来提高其性能，解决动态算法配置中的挑战。

DACBench is a standardized benchmark library designed to collect and standardize existing DAC benchmarks across various AI domains, while providing templates for new benchmarks. Co-developed by the Institute of Information Processing (tnt) and Leibniz University Hannover, this dataset includes six initial diverse DAC benchmarks covering fields such as AI planning, deep learning, and evolutionary computation. The design of DACBench prioritizes flexibility, reproducibility, extensibility, automated documentation and visualization. It has broad application scenarios, aiming to improve the performance of target algorithms by dynamically controlling their high-level parameter settings, and addressing the challenges in dynamic algorithm configuration.