Dataset for 'No-Tardiness Scheduling for Parallel Machine Workshop Based on Deep Reinforcement Learning'

This dataset supports the paper "No-Tardiness Scheduling for Parallel Machine Workshop Based on Deep Reinforcement Learning". It contains anonymized, real-world scheduling data from a home appliance manufacturing environment and is intended for benchmarking intelligent scheduling algorithms in dynamic production settings. The dataset includes: Basic information files (machines, products, sub-components, tooling, changeover matrices). Multiple demand instances for rescheduling experiments. Scheduling results from human planners, DRL (ASFF-VFA), GA, and hybrid (ASFF-VFA-GA) methods. Initial machine states and stock levels. All identifiers have been anonymized. Sensitive numerical values have been scaled, and time fields are expressed in seconds from a zero-based production start point. For field definitions, file formats, and usage instructions, please refer to the accompanying PDF: README.pdf . Please cite this dataset and the associated study if used in your research.本数据集配套论文《基于深度强化学习的并行机器车间无拖期调度》（No-Tardiness Scheduling for Parallel Machine Workshop Based on Deep Reinforcement Learning）。它收录了家电制造环境下的匿名化真实调度数据，旨在为动态生产场景中的智能调度算法提供基准测试支持。 本数据集涵盖以下内容： 1. 基础信息文件（包含机器、产品、零部件、工装夹具、切换矩阵）； 2. 多组用于重调度实验的需求实例； 3. 人工调度员、深度强化学习（Deep Reinforcement Learning，DRL，采用ASFF-VFA框架）、遗传算法（Genetic Algorithm，GA）以及混合算法（ASFF-VFA-GA）生成的调度结果； 4. 初始机器状态与库存水平。 所有标识符均已完成匿名化处理，敏感数值已做缩放处理，时间字段以生产启动时刻为零点，以秒为单位进行计量。 关于字段定义、文件格式与使用说明，请参阅随附的PDF文档README.pdf。 若您的研究中使用本数据集，请一并引用该数据集及其关联研究。