LINC-BIT/AirCa

Contents

• —1. About Dataset
• —2. Download
• —3. Description
• —3.1 AirCa-W
• —3.2 AirCa-N
• —3.3 Constraints description
• —4. The AirCa APIs
• —5. References <!-- - 7. Citations -->

<!-- --- license: apache-2.0 tags:

• —logistic
• —Spatial-Temporal
• —cargo-loading --- -->

Dataset Download: https://huggingface.co/datasets/LINC-BIT/AirCa Dataset Website: https://huggingface.co/datasets/LINC-BIT/AirCa Code Link: https://github.com/LINC-BIT/AirCa Paper Link:

1. About Dataset

AirCa is a publicly available aircraft cargo loading dataset with millions of instances from industry. It has three unique characteristics: (1) Large-scale, AirCa contains in total 6,071k records and 1,092k flights, covering 6 aircraft types and 425 airports over a total span of 9 months. (2) Comprehensive information, AirCa is delivered to provide rich information pertaining to aircraft cargo loading, including detailed cargo characteristic information, loading-event logs, flight destination, and comprehensive loading constraints in practical scenarios. (3) Diversity, AirCa aims to increase data diversity from three perspectives: destination diversity, Flight diversity, and Constraint diversity.

This figure illustrates an air cargo loading scenario comprising three parts: (1) Air cargo has two types: bulk cargo, which consists of individual sub-cargoes, and Unit Load Devices (ULDs), which are pre-packed standardized containers. (2) Cargo holds accommodate bulk cargoes (e.g., narrow-body aircraft such as the A320) and ULDs (e.g., wide-body aircraft such as the B777). Notably, when loaded into bulk cargo holds, different cargo detaching granularities produce a massive number of loading options. (3) Constraints determine the feasibility of loading operations, including cargo constraints, cargo hold constraints, and loading operation constraints. In this figure's example, the flight has two destinations, and hence the ULDs are categorized into two segments. Such multi-segment cargo loading further complicates the combinatorial optimization problem.

2. Download

AirCa can be used for research purposes. Before you download the dataset, please read these terms. Then put the data into "./data/raw/". The structure of "./data/raw/" should be like:

* ./data/raw/  
    * split_by_aircraft_type    
        * A320.csv   
        * ...    
    * split_by_date  
        * BAKFLGITH_LOADDATA2024-10-12.csv  
        * ...

import pandas as pd
>>> import pandas as pd
>>> df = pd.read_csv("BAKFLGITH_LOADDATA2024-10-12.csv")
>>> df.head(3)
       FLIGHT  TYPE DEST  WEIGHT  ... CONT PRIORITY VOLUME  SPECIAL CARGO
0  3744617311  A320  SIN     177  ...  NaN        1    0.0            NaN
1  3744617311  A320  SIN     177  ...  NaN        1    0.0            NaN
2  3744617332  A320  SIN     560  ...  NaN        1    0.0            NaN

3. Description

Below is the detailed field of each sub-dataset.

3.1 AirCa-W

3.2 AirCa-N

3.3 Constraints description

4. The AirCa APIs

In addition to our AirCa dataset, we release the AirCa package, including three types of APIs. It is designed to faciliate researchers in developing aircraft cargo loading applications.The details are presented as follows:

DataDownloader. This API allows researchers to download the AirCa data. the code presents how to utilize the DataDownloader API to download the up-to-date AirCa data. DataDownloader. This API allows researchers to download the AirCa data. Figure 4 presents how to utilize the DataDownloader API to download the up-to-date AirCa data.

from api . download_airca import AirCaDownloader
downloader = AirCaDownloader ()
# Download data A320
downloader . download_AirCa ( url , path , aircraft_type =" A320 " ,
date =" 2024 -10 -12 ")
# Download data B737
downloader . download_AirCa ( url , path , aircraft_type =" B737 " ,
date = None )
# Download data for all available aircraft types
downloader . download_AirCa ( url , path , aircraft_type = None ,
date = None )

DataRetriever. This API enables researchers to conveniently obtain the AirCa data stroed in the local machine. For instance, the code shows how to employ the DataRetriever API to obtain the AirCa data for aircraft type B777.

from api . retriever import Retriever
retriever = Retriever ()
# Enter the type A320
retriever . retrieve ( path = path , aircraft_type = " A320 ")
# Enter the type B777
retriever . retrieve ( path = path , aircraft_type = " B777 ")
# Enter the type B787
retriever . retrieve ( path = path , aircraft_type = " B787 ")

DataLoader. This API is designed to assist researchers in their applications of aircraft cargo loading. It allows researchers to flex- ibly and seamlessly merge multiple modalities of AirCa data. It exposes the AirCa through a DataLoader object after performing necessary data preprocessing techniques. A PyTorch example of using our DataLoader API for training DNNs is shown in the code.

import torch
from torch . utils . data import DataLoader
# generate AirCa ( A320 ) dataset for training
dataloader1 = DataLoader ( AircraftDataset ( path ," A320 ") ,
batch_size = batch_size , shuffle = True )
# generate AirCa ( B777 ) dataset for training
dataloader2 = DataLoader ( AircraftDataset ( path ," B777 ") ,
batch_size = batch_size , shuffle = True )
# generate AirCa ( B787 ) dataset for training
dataloader3 = DataLoader ( AircraftDataset ( path ," B787 ") ,
batch_size = batch_size , shuffle = True )
train_model ( dataloader1 , baseline_name , criterion ,
optimizer , epochs =600)

5. References

[1] Zhao, X., Dong, Y., & Zuo, L. (2023). A combinatorial optimization approach for air cargo palletization and aircraft loading. Mathematics, 11(13), 2798. [2] Mesquita, A. C. P., & Sanches, C. A. A. (2024). Air cargo load and route planning in pickup and delivery operations. Expert Systems with Applications, 249, 123711. [3] Yan, S., Lo, C.-T., & Shih, Y.-L. (2006). Cargo container loading plan model and solution method for international air express carriers. Transportation Planning and Technology, 29(6), 445–470. [4] Yan, S., Shih, Y.-L., & Shiao, F.-Y. (2008). Optimal cargo container loading plans under stochastic demands for air express carriers. Transportation Research Part E: Logistics and Transportation Review, 44(3), 555–575. [5] Limbourg, S., Schyns, M., & Laporte, G. (2012). Automatic aircraft cargo load planning. Journal of the Operational Research Society, 63(9), 1271–1283. [6] Zhao, X., Yuan, Y., Dong, Y., & Zhao, R. (2021). Optimization approach to the aircraft weight and balance problem with the centre of gravity envelope constraints. IET Intelligent Transport Systems, 15(10), 1269–1286. [7] Lurkin, V., & Schyns, M. (2015). The airline container loading problem with pickup and delivery. European Journal of Operational Research, 244(3), 955–965. [8] Zhu, L., Wu, Y., Smith, H., & Luo, J. (2023). Optimisation of containerised air cargo forwarding plans considering a hub consolidation process with cargo loading. Journal of the Operational Research Society, 74(3), 777–796. [9] Chenguang, Y., Liu, H., & Yuan, G. (2018). Load planning of transport aircraft based on hybrid genetic algorithm. In MATEC’18, Vol. 179 (pp. 01007). EDP Sciences. [10] Dahmani, N., & Krichen, S. (2016). Solving a load balancing problem with a multi-objective particle swarm optimisation approach: application to aircraft cargo transportation. International Journal of Operational Research, 27(1-2), 62–84. [11] Dahmani, N., & Krichen, S. (2013). On solving the bi-objective aircraft cargo loading problem. In ICMSAO’13 (pp. 1–6). IEEE. [12] Gajda, M., Trivella, A., Mansini, R., & Pisinger, D. (2022). An optimization approach for a complex real-life container loading problem. Omega, 107, 102559.

<!-- # 7. Citation --> <!-- If you find this helpful, please cite our paper:

@misc{wu2023lade,
      title={LaDe: The First Comprehensive Last-mile Delivery Dataset from Industry}, 
      author={Lixia Wu and Haomin Wen and Haoyuan Hu and Xiaowei Mao and Yutong Xia and Ergang Shan and Jianbin Zhen and Junhong Lou and Yuxuan Liang and Liuqing Yang and Roger Zimmermann and Youfang Lin and Huaiyu Wan},
      year={2023},
      eprint={2306.10675},
      archivePrefix={arXiv},
      primaryClass={cs.DB}
}