logistics

该数据集是一个经典的规划领域，目标是通过飞机和卡车将一系列包裹在不同地点和城市间进行搬运。生成的问题不仅有效、多样，而且难度较大，与特定领域生成器产生的问题相比具有明显优势。该数据集涵盖了包含15至40个原子的各类问题，任务涉及规划和问题生成及解决。

This dataset is a classic planning domain where the goal is to transport a series of packages between various locations and cities using airplanes and trucks. The generated problems are not only valid, diverse and challenging, but also have significant advantages over those produced by domain-specific generators. This dataset covers various problems with 15 to 40 atoms, and the tasks involve planning, problem generation and problem solving.

搜集汇总

背景与挑战

背景概述

该数据集是'S-PlanningProblemGeneration'项目的一部分，专注于使用神经符号方法（NeSIG）自动生成PDDL规划问题，确保问题有效、多样且难以解决。数据集包含多个PDDL领域（如blocksworld、logistics和sokoban）的代码和实验数据，支持训练和测试问题生成，并提供了详细的安装和使用指南。