Application and Development of Artificial Intelligence in Emergency

Logistics: Taking Earthquake Rescue as an Example

Chao Liu

and Yixiang Zhu

School of Economics and Management, Hefei University, Hefei ,230601, China

School of Management Engineering, Xuzhou University of Technology, Xuzhou, 221111, China

Keywords: Emergency Logistics, Earthquake, Artificial Intelligence, Internet of Things, Cloud Computing.

Abstract: Since the beginning of the 21st century, to better ensure the livelihood of residents in earthquake-stricken

areas, how to improve the efficiency of the emergency logistics system has become an increasingly important

issue in the academic circle. In this paper, artificial intelligence algorithms, cloud computing, Internet of

Things and other technologies in earthquake rescue cases in different regions are studied. While fusing various

schemes, the technical route of this research always adheres to the principle of people-oriented. The

conclusion of this study indicates that, with adequate information guarantee and appropriate selection of

schemes, the benefits brought by new technologies such as artificial intelligence to emergency logistics are

very significant.The significance of this research lies in providing a practical and technologically advanced

framework for enhancing emergency response capabilities, thereby minimizing human and economic losses

during disasters. Furthermore, it offers valuable insights for policymakers and disaster management agencies

to optimize resource allocation and improve coordination efficiency in future earthquake rescue operations.

Additionally, this study underscores the transformative potential of integrating cutting-edge technologies like

AI and IoT into humanitarian logistics, setting a benchmark for innovation-driven disaster resilience strategies.

1 INTRODUCTION

Since the dawn of the 21st century, human society has

been rapidly advancing while concurrently

confronting the repercussions of natural disasters. A

crucial practical challenge that has emerged is how to

safeguard the livelihoods of residents in disaster

zones during such events, particularly with regard to

leveraging emerging technologies to enhance the

efficiency of emergency logistics.

Looking back on history, during the early

exploration phase spanning from the mid-20th

century to the late 20th century, computers were

initially utilized for processing seismic data, yet they

were not effectively integrated. Subsequently, in the

sluggish development phase from the early 21st

century to 2010, emphasis was placed on the

collection and organization of earthquake-related

data, and basic information management systems

started to be implemented in earthquake emergency

responses. However, the application of artificial

intelligence (AI) technology in emergency logistics

Corresponding author

was relatively limited. Moving into the rapid

development phase from 2010 to 2020, scholars

began to focus on optimizing earthquake emergency

logistics through AI. Gradually, AI technology was

introduced into the field of earthquake emergencies,

albeit mostly remaining at the stages of theoretical

research and experimentation.

Currently, as technologies such as deep learning,

big data, and the Internet of Things continue to evolve

and converge, artificial intelligence (AI) is

increasingly integrating with GIS, GPS, IoT, big data,

and other technologies. Consequently, intelligent

decision support systems are also maturing, and AI is

offering more robust technical support for various

aspects of earthquake emergency logistics.

Nevertheless, emergency logistics still encounters

numerous challenges, with the most pressing issue

being high costs and low efficiency, necessitating

urgent measures to reduce costs and enhance

efficiency.

Specifically, at present, there have been many

research and achievements in earthquake and

Liu, C. and Zhu, Y.

Application and Development of Artiﬁcial Intelligence in Emergency Logistics: Taking Earthquake Rescue as an Example.

DOI: 10.5220/0014354900004718

Paper published under CC license (CC BY-NC-ND 4.0)

In Proceedings of the 2nd International Conference on Engineering Management, Information Technology and Intelligence (EMITI 2025), pages 343-347

ISBN: 978-989-758-792-4

343

emergency logistics, but there is still room for

discussion in the cross-field of artificial intelligence.

Therefore, the main content of this paper is to focus

on agility and promote the integrated application of

artificial intelligence and traditional emergency

logistics. In short, it is to comprehensively analyse the

specific situation of earthquake rescue cases in

different regions, adhere to the people-oriented

concept, better integrate traditional emergency

logistics with new technologies such as artificial

intelligence, cloud computing and Internet of things,

and improve the comprehensive ability of emergency

logistics, so as to better respond to earthquake

disasters.

The purpose of this study can be divided into two

aspects: Theory and reality. In terms of theory, the

research aims to fill the gap between the theoretical

basis and the target field and lay the foundation for

subsequent research. In terms of reality, we hope to

help social organizations, enterprises and institutions,

and organs better apply emerging technologies such

as artificial intelligence to the field of emergency

logistics, so as to better respond to frequent natural

disasters, improve rescue capabilities, and protect

people's lives and property.

2 CASE ANALYSIS

The core viewpoint of this article is that "the

application and development of artificial intelligence

in emergency logistics can be reflected in different

regions and environments." According to the

characteristics of different regions, people can

combine artificial intelligence and other new

technologies with traditional emergency logistics to

reduce costs and increase efficiency.

Cost reduction and efficiency improvement can be

broken down into cost reduction and efficiency

enhancement.

In order to improve efficiency, this study found

the following commonalities after interpreting

different literature: First, the level of data acquisition

and analysis, and evaluation in disaster areas is

limited, which makes it difficult to formulate plans.

Second, the decision-making strategy is not perfect

enough and performs poorly when facing an uncertain

environment. Third, problems such as insufficient

transportation capacity are caused by limited

equipment selection and insufficient development of

usage methods.

First of all, an algorithm suitable for the

mobilization of emergency supplies is necessary. Li's

research proposes an emergency logistics resource

scheduling algorithm in the cloud computing

environment, called Emergency Supplier Distribution

Mobile Edge Computing (ESD-MEC), to address the

dual constraints of where to store emergency logistics

equipment and where to dispatch rescue vehicles(Li,

2024). This method uses road transportation for local

delivery. Suppliers can obtain information and

manage data in the emergency supplier management

system. The cloud computing platform allocates and

calculates a large amount of data in densely populated

areas through multiple MEC servers and uses these to

strengthen the connection among all emergency

logistics network nodes and reduce costs, minimizing

time expenditure and transportation costs. With the

algorithm, practical data is needed. The research of

Ding et al. proposed an earthquake emergency

logistics technology system based on Internet of

Things (IoT) technology. It can realize the systematic

application of Internet of Things technology in all

stages of earthquake rescue (Ding, Zhao and Li,

2020). In this way, various devices can better obtain

and exchange information with the help of the

Internet of Things. For example, through RFID

technology, inventory can be visualized, and when

performing allocation tasks, it can be quickly located

and accessed, facilitating subsequent management

and allocation. Sorting out the information is as

important as conducting a preliminary analysis. The

research of Cao proposes to achieve intelligent and

flexible disaster management by using AI for Smart

Disaster Resilience (AISDR) technology(Cao, 2023).

For example, Remote Sensing (RS) technology

(satellites, unmanned aerial vehicles, etc. can be used

as shooting equipment) and combined with

autonomous visual recognition AI models to analyse

the disaster situation in the disaster-stricken area,

such as water pollution, air pollution, damage to

ground buildings and road conditions, etc. When the

situation in the disaster-stricken area is not fully

understood, RS technology can make up for some

deficiencies and help decision-makers sort out the

complex information as much as possible to assist in

making wise decisions. As for decision-making,

although traditional centralized systems can process

all information centrally, they may not respond

promptly in the face of unexpected situations due to

overly long chains. The research of Han proposed the

Distributed Vehicle Routing Algorithm (DVRA),

each vehicle in the system has the ability to collect,

identify information and make decisions, meeting

approximately 70.7% of the demand and achieving

about 95% of the optimal solution target. Compared

with the centralized system, this distributed system

can better cope with the uncertainties on site(Han and

EMITI 2025 - International Conference on Engineering Management, Information Technology and Intelligence

344

Jeong, 2025). This technology not only makes the

means of information acquisition more abundant, but

also decentralizes the decision-making power, which

greatly reduces the possibility of losses caused by the

failure of the centralized system to respond to sudden

changes. Since the decision is made, appropriate

equipment needs to be selected to perform the task.

For example, if materials are transported only by

vehicle, they will not only be subject to road

conditions but also make the management of

materials of different emergency degrees and natures

chaotic and unable to be sent to the needed areas in

time, resulting in irreparable consequences. Edwards

et al. 's research mentioned that using drones to

transport light and small emergency supplies such as

vaccines and water purification tablets can

compensate for part of the ground capacity when it is

insufficient, reducing the pressure on ground work

(Edwards, Subramanian and Chaudhuri et al, 2024).

In this way, ground transportation capacity such as

vehicles can concentrate on transporting large and

heavy materials such as food, tents, emergency

furniture, and mechanical equipment, which is

expected to reduce the possibility of casualties caused

by the untimely transportation of emergency

materials, and lower the damage rate and distribution

difficulty of materials during transportation due to the

mixture of sizes and types. In addition, it is also very

important to prepare a method that can improve the

deployment efficiency of emergency materials and

optimize the routes. Yang et al. 's research proposed

Optimization of a Two-Stage Emergency Logistics

System considering public psychological risk

perception under earthquake disaster. Particle Swarm

Optimization (PSO) is used to improve the Sparrow

Search Algorithm (SSA) to further solve the model

(Yang and Zhang et al, 2024). This study takes public

psychological risks into account. Considering the

urgency and the inability to complete the optimal

deployment in the first place, the method of deploying

first and then optimizing was adopted, which

minimized the psychological and physiological losses

of the affected people.

In terms of cost reduction, it might be that the

rescue workers made a wrong prediction of the

demand in the disaster-stricken area, resulting in the

input of the wrong types or quantities of resources and

causing waste. It could also be due to improper site

selection or excessive reliance on government rescue,

resulting in excessively high costs. The research of

Lin et al. mentioned a method for earthquake

emergency medicines based on Gradient-Boosting

Decision Trees （ GBDTs) and Attention-Free

Transformer and Long Short Term（AFT-LSTM),

this model achieved an average absolute percentage

error of 1.67% predicted by the benchmark test, an

average square error of 4.6%, and a square correlation

coefficient of 0.96, which are highly consistent with

the actual number of affected people(Lin,Yan and

Zhang et al, 2025). This model enables emergency

relief supplies to be better distributed with the support

of mathematical methods, reducing waste. Geng et al.

‘s research proposed an optimization plan for

warehouse location selection and material allocation,

considering the perception of disaster victims' pain

under the collaborative rescue efforts of the

government and enterprises. While reducing costs

and increasing efficiency, it introduced the

sustainable development concept of "people-

oriented" into disaster relief work (Geng and Hou,

2021), reducing the waste of time and energy in

transportation caused by improper location selection

and lowering the possibility of excessive reliance on

government rescue. Besides, the Post-Earthquake

Emergency based on a multi-objective genetic

algorithm based on adaptive large neighbourhood

search proposed by Pu et al, Logistics Location-

Routing Optimization Considering Vehicle Three-

Dimensional Loading Constraints, in 20 instances,

compared with multi-objective evolutionary

algorithms, etc. It was significantly superior in 17 and

16 instances respectively. When averaging the mean

and variance of 20 runs, this algorithm shows a larger

average mean and a smaller average variance,

indicating that its performance is superior to that of

the comparison algorithm (Pu and Zhao, 2024). This

plan can incorporate practical factors such as the

physical conditions of vehicles into the site selection

considerations, and it is also a very important part of

cost reduction and efficiency improvement.

As for evaluating the implementation of the plan,

the amount of loss is one of the important indicators

to judge the success of this rescue. Akter et al. 's

research mentioned the monthly and annual average

nighttime light data collected through the visible and

infrared imaging radiometer suite instrument as an

alternative method for capturing non-economic

welfare losses (Akter, Chairunissa and Pundit, 2024).

At the same time, the assessment should also be

combined with methods including but not limited to

continuing to compare the living standards of the

people in the disaster-stricken area and the

establishment of the emergency logistics system

before and after receiving disaster relief through

objective detection equipment (such as visual

recognition models + optical images for analysis),

comparing with other similar disaster-stricken areas,

and randomly interviewing the people in the disaster-

Application and Development of Artiﬁcial Intelligence in Emergency Logistics: Taking Earthquake Rescue as an Example

345

stricken area. Such a series of operations can yield a

relatively more objective and accurate actual effect of

the plan implementation compared to a single

method.

Certainly, when it comes to a specific rescue

operation, it cannot be mechanically applied. Instead,

it should be discussed in a classified manner based on

the actual situation. For example, if there is a road

available, use a car; if there is no road available, use

a boat. When there is sufficient information, the route

and materials are directly formulated without

assessment; when there is insufficient information,

assessment is conducted first and then the relevant

plans are formulated. When the transportation

capacity is sufficient, cost should be considered;

when the capacity is insufficient, emergency supplies

should be given priority.

3 ADVISE

Although the problems that emergency logistics

needs to deal with are complex, they are often

universal, so some suggestions are interchangeable.

3.1 Imformation Supporty

In order to make effective decisions, emergency

logistics needs to obtain sufficient information

support. This can be divided into two aspects:

sufficient information and accurate information

judgment. On the one hand, in order to obtain

sufficient information, it is not only necessary to

make full use of emerging technologies such as the

Internet of things, cloud computing and unmanned

aerial vehicles, expand the means and perspectives of

obtaining information, and improve the ability to

obtain information, so as to obtain more direct

information. Moreover, it is necessary to give full

play to the analysis ability of artificial intelligence to

receive and screen relevant information obtained by

traditional methods such as the Internet and offline

communities to the greatest extent, so as to better

eliminate the false and retain the true, and improve

the ability and accuracy of obtaining indirect

information. On the other hand, to make accurate

information judgment, we should reasonably use

artificial intelligence technology on the basis of

existing information, establish big data model, and

then analyse existing information combined with the

basic situation of earthquake disaster, improve the

accuracy of demand prediction, provide support for

realistic decision-making, so as to make better

decisions, speed up the efficiency of emergency

logistics system operation, reduce unnecessary waste

of resources, and achieve efficient operation.

3.2 People Oriented

During the operation of emergency logistics,

personnel safety should be taken as the highest

criterion, giving priority to the effect of rescue, and

then considering the cost. On the one hand, this

requires giving priority to saving lives and protecting

survivors. In short, on the basis of rapid and scientific

needs assessment, it comprehensively considers the

concerns of vulnerable groups such as the elderly, the

weak, the sick, the disabled and pregnant, and

provides safe and reliable guarantee materials for the

disaster area at the fastest speed, so as to reduce the

chance of their secondary injury as much as possible

and ensure the life safety of the affected people to the

greatest extent. On the other hand, it also requires

comprehensive consideration of the psychological

care for the victims. To put it simply, on the basis of

ensuring the supply of materials, we should respect

the local cultural customs, give full consideration to

the mental health of the victims, provide them with

certain psychological counselling service, reduce

their psychological pressure and trauma, and enable

them to better come out of the shadow of the disaster.

In addition, the materials for disaster relief are not

unlimited, which requires the rational use of AI

algorithm on the basis of ensuring the supply of

materials, further optimizing the logistics location

and logistics route, reducing costs, improving

efficiency, and preserving strength for future

recovery and reconstruction.

3.3 Feedback Optimization

Emergency logistics should have perfect means to

obtain feedback in order to continue to improve. On

the one hand, it needs to expand the sources of

feedback and improve the multi-level and

comprehensive feedback network. Specifically, it is

not only possible to recruit local volunteers on the

basis of setting up message areas and manual

information collection points, conduct regular visits

to survivors and rescue related personnel, collect the

needs of different regions and different groups, and

summarize and report them, so as to form a perfect

offline network. In addition, a simple feedback

website can be established and a smooth feedback

telephone can be provided, so that relevant personnel

can carry out online feedback in a timely and efficient

manner. On the other hand, it needs to ensure the

efficiency of feedback optimization. Specifically, we

EMITI 2025 - International Conference on Engineering Management, Information Technology and Intelligence

346

should not only establish and improve the tracking

mechanism, submit the feedback to the relevant

system, and the system will estimate the time required

according to the specific situation, and automatically

track the processing progress. Moreover, on the

premise of protecting the privacy of relevant

personnel and allowing the actual conditions, we can

make full use of the Internet and artificial

intelligence, comprehensively analyse the

suggestions of external personnel and the opinions of

internal personnel, optimize the logistics process as

soon as possible, improve deficiencies and improve

efficiency.

4 CONCLUSIONS

Through research, this paper found that the impact of

the earthquake is diverse. Emergency logistics not

only needs to deal with the economic and non-

economic impact of the earthquake and its secondary

disasters, but also needs to consider the psychological

impact of equipment, roads, personnel and other

complex issues. Therefore, in order to better apply

artificial intelligence to the field of emergency

logistics, we should combine artificial intelligence to

comprehensively consider various problems caused

by the earthquake, establish a model suitable for

emergency material mobilization, optimize the

demand forecasting method, and implement and

evaluate the specific situation of the scheme.

The contribution of this paper is to make up for the

gap in improving the efficiency of emergency

logistics by organically integrating traditional

emergency logistics with new technologies such as

artificial intelligence, cloud computing and the

Internet of things, focusing on agility and adhering to

the people-oriented principle, which is conducive to

the application of artificial intelligence technology to

the field of emergency logistics by the government

and relevant organizations, and better protect the

safety of people's lives and property. There are still

some problems in the current research, such as the

imperfect model of emergency material mobilization

and the incomplete analysis of the specific situation

of earthquakes. Future research should further

improve the emergency logistics processing model

and improve the applicability and accuracy of the

processing, combined with the relevant data of more

earthquake-prone regions.

AUTHORS CONTRIBUTION

All the authors contributed equally and their names

were listed in alphabetical order.

REFERENCES

Akter, S., Chairunissa, T. F., & Pundit, M. (2024).

Assessing "Invisible Loss" during medium-term

earthquake recovery: The case of Indonesia's 2016 Aceh

earthquakes. Economic Disasters and Climate Change,

8, 541-561.

Cao, L. (2023). AI and data science for smart emergency,

crisis and disaster resilience. Data Science and Analytics,

15, 231-246.

Ding, L., Zhao, L., Li, L., et al. (2020). Study of emergency

logistics technology system for earthquake equipment

based on Internet of Things. Journal of Catastrophology,

35(2), 200-205.

Edwards, D., Subramanian, N., Chaudhuri, A., et al. (2024).

Use of delivery drones for humanitarian operations:

Analysis of adoption barriers among logistics service

providers from the technology acceptance model

perspective. Annals of Operations Research, 335, 1645-

1667.

Geng, J., Hou, H., & Geng, S. (2021). Optimization of

warehouse location and supplies allocation for

emergency rescue under joint Government-Enterprise

cooperation considering disaster victims distress

perception. Sustainability, 13(19), 10560.

Han, S., & Jeong, H. (2025). Research on efficient relief

logistics planning under massive natural disasters.

Scientific Reports, 15, 11595.

Li, T. (2024). Emergency logistics resource scheduling

algorithm in cloud computing environment. Physical

Communication, 102340.

Lin, Y., Yan, W., Zhang, Y., Zhang, H., Zhang, H., & Wang,

D. (2025). Emergency drug demand forecasting in

earthquakes with XGBoost and AFT-LSTM.

Sustainability, 17(5), 1910.

Pu, X., & Zhao, X. (2024). Post-earthquake emergency

logistics location-routing optimization considering

vehicle three-dimensional loading constraints.

Symmetry, 16(8), 1080.

Yang, H., Zhang, P., Zhang, P., et al. (2024). Optimization

of a two-stage emergency logistics system considering

public psychological risk perception under earthquake

disaster. Scientific Reports, 14, 31983.

Application and Development of Artiﬁcial Intelligence in Emergency Logistics: Taking Earthquake Rescue as an Example

347