five

Data for: Permutation Flow Shop Scheduling with Multiple Lines and Demand Plans Using Reinforcement Learning

收藏
DataCite Commons2025-05-01 更新2025-05-17 收录
下载链接:
https://data.mendeley.com/datasets/5txxwj2g6b
下载链接
链接失效反馈
官方服务:
资源简介:
Description of dataset This repository contains the datasets used in our study "Permutation Flow Shop Scheduling with Multiple Lines and Demand Plans Using Reinforcement Learning". The repository provides three datasets. The main dataset in the folder data contains 1050 problem instances for the multi-line permutation flow shop problem. The dataset extends the dataset introduced by Taillard (1993) to multiple production lines and demand plans. The processing times were sampled from the uniform interval [1,99]. In addition, we sampled the demand plans from a multinomial distribution with equal probability for each job type, which leads to rather balanced demand plans. The folders data_lin and data_exp each contain 150 problem instances with more imbalanced demand plans, where the probabilities of each job type decrease linearly or exponentially. However, the quantity to be produced for each job type is greater than zero. Dataset structure Each dataset of the main study is structured in 15 subfolders. Each folder contains problem instances for a combination of layout and processing time variation. Folder name notation: Tai_PFSP_L_<b> A: Number of production lines (1-3) B: Processing time variation (1-5) All of these folders contain 70 problem instances. A problem file is a combination of the problem layout (sequence length, number of machines and stations) and the demand plan variation. The processing times are fixed for one problem characteristic, but the ten demand plans are different. File name notation: t____.mix C: Number of production lines (1-3) D: Identifier of the layout (1-7) E: Sequence length (20,100,500) F: Number of machines per line (5,10,20) G: Number of job types (5,10,20) H: Number of demand plan variation (1-10) Each file represents a different problem in text format with the following notation: Line 1: Demand plan Line 2: Layout Type Line 3: Number of machines Line 4: Number of machines per line Line 5: Number of total machines with synchronization machine Line 6: Number of job types Line 7-end: Processing times in matrix form for machines (rows) and job types (columns)</b>

数据集说明 本仓库包含我们在研究《基于强化学习(Reinforcement Learning)的多产线带需求计划置换流水车间调度(Permutation Flow Shop Scheduling with Multiple Lines and Demand Plans Using Reinforcement Learning)》中使用的数据集。本仓库共提供三类数据集。主数据集存放于data文件夹中,包含1050个针对多产线置换流水车间调度问题的问题实例。该数据集是对Taillard(1993)提出的经典数据集的扩展,适配了多条产线与需求计划场景。其工件加工时长从均匀区间[1,99]中采样得到。此外,我们针对每种工件类型以等概率从多项分布(multinomial distribution)中采样生成需求计划,以此得到分布较为均衡的需求计划集。 data_lin与data_exp文件夹各包含150个问题实例,其需求计划分布更不均衡:各类工件的出现概率分别呈线性或指数级递减,但每类工件的生产需求量均大于0。 数据集结构 本研究的每一类数据集均包含15个子文件夹,每个子文件夹对应一类产线布局与加工时长波动的组合场景。文件夹命名规则如下:Tai_PFSP_L_<B>,其中: A为产线数量(取值范围1-3), B为加工时长波动系数(取值范围1-5)。 每个子文件夹包含70个问题实例。 单个问题文件包含问题布局(序列长度、机床数量与工位配置)与需求计划波动的组合信息。针对同一类问题特征,加工时长固定,但包含10种不同的需求计划变体。 文件命名规则:t____.mix,其中各字段含义如下: C为产线数量(1-3), D为布局标识符(1-7), E为序列长度(20、100、500), F为单条产线的机床数量(5、10、20), G为工件类型数量(5、10、20), H为需求计划变体编号(1-10)。 每个文件均为文本格式的问题实例,其内容格式如下: 第1行:需求计划 第2行:布局类型 第3行:机床总数 第4行:单条产线的机床数量 第5行:包含同步机床的总机床数 第6行:工件类型数量 第7行及之后:以矩阵形式呈现的机床(行)与工件类型(列)对应的加工时长
提供机构:
Mendeley
创建时间:
2019-11-04
搜集汇总
数据集介绍
main_image_url
背景与挑战
背景概述
该数据集包含1050个多线置换流水车间调度问题的实例,扩展了经典Taillard数据集,增加了多生产线和需求计划的内容。数据集分为三个部分,分别对应不同需求计划平衡程度的问题实例,每个实例包含详细的布局、处理时间和需求计划信息。
以上内容由遇见数据集搜集并总结生成
5,000+
优质数据集
54 个
任务类型
进入经典数据集
二维码
社区交流群

面向社区/商业的数据集话题

二维码
科研交流群

面向高校/科研机构的开源数据集话题

数据驱动未来

携手共赢发展

商业合作