Data of Ship Block Scheduling in English

This study is based on the following assumptions: Assumption 1: The projection shape of each block position and block is a minimum bounding rectangle. Assumption 2: Each block position can process only one block at a time. Assumption 3: Labor force is sufficient, without absenteeism or other anomalies. Assumption 4: Blocks of different sizes can find suitable positions for processing. Assumption 5: Each block is completed at only one block position. Assumption 6: All block positions are empty at the beginning of scheduling. The dataset includes input data of blocks and block positions, scheduling results of EDDQN, DDQN, and five heuristic algorithms, as well as raw data for figures in the paper. A key observation is that the EDDQN algorithm shows stronger advantages in large‑scale scheduling problems than in small‑ and medium‑scale cases. The data reflects scheduling performance under the stated assumptions. Lower makespan, lower cost, and fewer position switches indicate better scheduling performance.Users can load the input data, run the provided algorithms, and reproduce the experimental results and figures in the paper. Folder Structure data_en/ ├─ Raw_input_ Data/ │ ├─ block_information_table.xlsx │ ├─ initial_block_position_information.xlsx │ └─ block_position_information.xlsx ├─ Result/ │ ├─ DDQN scheduling.xlsx │ ├─ EDDQN_scheduling_results.xlsx │ ├─ scheduling_results_heuristic_earliest_start_time_20250605_014601.xlsx │ ├─ scheduling_results_heuristic_longest_processing_time_20250605_014601.xlsx │ ├─ scheduling_results_heuristic_resource_utilization_based_20250605_014601.xlsx │ ├─ scheduling_results_heuristic_response_time_based_20250605_014601.xlsx │ └─ scheduling_results_heuristic_shortest_processing_time_20250605_014601.xlsx ├─ Figure6&7_raw_data/ │ ├─ DDQN_results.xlsx │ └─ EDDQN_results.xlsx ├─ Figure10_raw_data.xlsx └─ mapping_key_texts.csv 1) Input Data (`Raw_input_ Data/`) block_information_table.xlsx` Block-level production and scheduling attributes. Typical fields include: - block identifier / sequence, - vessel identifier and vessel type, - block type, - dimensions and weight, - processing time, - earliest/latest start time and latest completion time, - combination-related attributes (if applicable). Use:Main job-side input for scheduling models. #### `initial_block_position_information.xlsx` Initial assignment status between blocks and block positions. Typical fields include: - vessel/block identifiers, - earliest start time, - initial block-position description, - expected completion time. Use: Defines initial state and initialization constraints. block_position_information.xlsx Resource and spatial constraints for block positions. Typical fields include: - block-position ID and description, - area hierarchy, - size/capacity limits, - lifting capacity and height constraints, - labor-team attributes, - designated/general usage attributes.本研究基于以下假设前提： 1. 各加工工位（block position）与待加工分段（block）的投影轮廓均为最小外接矩形。 2. 每个加工工位单次仅可处理一个分段。 3. 劳动力储备充足，无缺勤或其他异常状况。 4. 不同尺寸的分段均可找到适配的加工工位开展作业。 5. 每个分段仅可在一个加工工位完成全部加工工序。 6. 调度任务初始时刻所有加工工位均处于空闲状态。 本数据集包含分段与加工工位的输入数据、EDDQN、DDQN以及五种启发式算法的调度结果，同时附带论文中所有插图的原始数据。 核心观测结果表明：相较于中小规模调度场景，EDDQN算法在大规模调度问题中展现出更显著的性能优势。 本数据集的调度性能结果均基于前述假设前提生成。总完工时间（makespan）更短、成本更低且工位切换次数更少，即代表更优的调度表现。用户可加载输入数据、运行内置算法，复现论文中的实验结果与插图数据。 ### 文件夹结构 data_en/ ├─ Raw_input_ Data/ │ ├─ block_information_table.xlsx │ ├─ initial_block_position_information.xlsx │ └─ block_position_information.xlsx ├─ Result/ │ ├─ DDQN scheduling.xlsx │ ├─ EDDQN_scheduling_results.xlsx │ ├─ scheduling_results_heuristic_earliest_start_time_20250605_014601.xlsx │ ├─ scheduling_results_heuristic_longest_processing_time_20250605_014601.xlsx │ ├─ scheduling_results_heuristic_resource_utilization_based_20250605_014601.xlsx │ ├─ scheduling_results_heuristic_response_time_based_20250605_014601.xlsx │ └─ scheduling_results_heuristic_shortest_processing_time_20250605_014601.xlsx ├─ Figure6&7_raw_data/ │ ├─ DDQN_results.xlsx │ └─ EDDQN_results.xlsx ├─ Figure10_raw_data.xlsx └─ mapping_key_texts.csv 1) 输入数据（`Raw_input_ Data/`） #### `block_information_table.xlsx` 存储分段（block）级别的生产与调度属性。典型字段包括： - 分段标识/加工序列 - 船舶标识与船舶类型 - 分段类型 - 尺寸与重量 - 加工时长 - 最早/最晚开始时间与最晚完工时间 - 组合关联属性（如适用） 用途：调度模型的核心作业侧输入数据。 #### `initial_block_position_information.xlsx` 存储分段（block）与加工工位（block position）的初始分配状态。典型字段包括： - 船舶/分段标识 - 最早开始时间 - 初始分段-工位分配关系 - 预期完工时间 用途：定义调度初始状态与初始化约束条件。 #### `block_position_information.xlsx` 存储加工工位（block position）的资源与空间约束信息。典型字段包括： - 加工工位ID与描述 - 区域层级 - 尺寸/容量限制 - 起吊能力与高度约束 - 作业班组属性 - 指定/通用使用属性。