Alibaba-NLP/Open-Travel

Open-Travel数据集是一个专注于旅行规划领域的数据集，旨在帮助用户完成行程规划的子任务。这些任务强调多约束推理、多工具协调以及结合用户特定约束（如预算限制、时间窗口、旅行团队和偏好配置文件）的个性化偏好。数据集包含1,626个用于强化学习（RL）的训练样本和250个测试样本，测试样本均匀分布在五个不同的子任务中：附近兴趣点搜索、多途径点路线规划、交通方式比较、单日城市旅行规划和多日旅行规划（泛化任务）。

The Open-Travel dataset is focused on the travel planning domain, designed to assist users in accomplishing itinerary planning subtasks. These tasks emphasize multi-constraint reasoning, multi-tool coordination, and personalized preferences intertwined with user-specific constraints (e.g., budget limits, time windows, traveling parties, and preference profiles). The dataset includes 1,626 RL training samples and 250 test samples, evenly distributed across five distinct subtasks: nearby point-of-interest (POI) search, route planning with multiple specified waypoints, transportation-mode comparison, one-day trip planning in a single city, and multi-day trip planning (Generalization task).