Ma Yu-xue Wu Zhi-feng Yang Xiao-fan
Qingdao University of Tchnology Shan Dong Qing Dao 266520
Abstract：Do you know what people's lives are like after the hurricane? People after the hurricane urgently need medical supplies. At the same time，we need to know the road traffic conditions in the disaster area in order to deliver medical supplies to the destination in time. In this paper，we establish disaster response system model and cargo container location model so on to solve these problems. And our efforts in building these models followed the path below.
Keywords：Hurricane rescue；Drone delivery；Programming model
The introduction
Firstly，we establish a disaster response system model by using the pruning algorithm to solve the composition of the drone fleet which consists of one D-type Drone and two E-type Drones. To simplify the problem，we choose a set of medical packages from five hospitals’ medical packages. Moreover，we put three cargo containers into three port hospitals including Fajardo，San Juan，Arecibo. According longitude and latitude calculation formula，we find that San Juan is nearer to San Pablo and Bayamon than Fajardo and Arecibo. For providing the main road conditions，we decide that the Drone E of San Juan performs medical supply delivery and video reconnaissance，and two remaining Drones only are able to detect roads.
Secondly，in order to minimize any need for buffer materials for unused space we establish the cargo container configuration model. The results are that the first cargo container loaded with a D-type Drone and 3834 MED1 and 1 MED2 docks at Arecibo；the second cargo container loaded with a Drone E and 2435 MED1，974 MED2，1458 MED3 docks at San Juan；the third cargo container loaded with a drone E and 2273 MED1，and 2274 MED3 docks at Fajardo. Besides，we list four transportation routes，and we chose the third route with the least time and distance to use . So far，we solve the shipping problem.
Thirdly，in order to get Drones video reconnaissance routes，we choose sixty cities across the major highways and roads by map of Puerto Rico. When Drones fly along these cities，they can provide road detection videos for HELP，Inc.，giving people the opportunity to provide adequate and timely response during or after natural disasters. We draw the figure of three video reconnaissance road maps by MATLAB with Floyd algorithm. When Drones fly from the starting point along their respective routes，the maximum range of road conditions can be detected. Besides，we get the first cargo container available to Arecibo hospital for 3834 days；the second cargo container available to San Juan hospital for 486 days and the third cargo container available to Fajardo hospital for 2273 days.
Finally，based on our model analysis and conclusions，we provide some proposals for the HELP，Inc. CEO about Drone’s functions recommendations. Furthermore，we evaluate our models from different aspects and make further improvements.
Key words：Floyd algorithm；pruning algorithm；disaster response system model，cargo container location model，aerial disaster relief response system
Introduction
1.1 Background
Hurricane is one of the main meteorological disaster. Every year from June 1st to November 30th is the American Atlantic hurricane season，and August to October is the peak period of hurricane activity. In recent years，hurricane always attacks American. Hurricane is usually accompanied by strong winds and heavy rains，meanwhile it threats people’s lives and property seriously，causing great impact on people’s livelihood，agriculture，and economy. So it is a natural disaster with great impact and serious damage. It is especially important for us to take active measures to do disaster relief work. From various documents 
at home and abroad，they are actively exploring various effective methods to enable relief supplies to reach the disaster area quickly.
1.2 Restatement of the Problem
Hurricane is kind of Tropical storm with heavy rainfall，in 2017，Puerto Rico suffered the attack of hurricane. It is urgent for victims to get emergency medical package in time，and the government needs to know the traffic situation in disaster area. Therefore，HELP，Inc（one NGO）is attempting to improve its response capabilities by designing a transportable disaster response system called ”DroneGo”. First，we load Drone Cargo Bay into shipping containers. Then we use cargo containers to transport shipping containers and emergency medical package to a particular disaster area that we choose. Finally，we use Drones send emergency medical packages to five delivery locations.
We regard the number of five hospitals’ emergency medical packages as standard，so the of emergency medical package a day is a constant. By pruning programming
，we can calculate the Drone mode required to transport the medical packages，and we can establish a linear optimization model to get the number of Drone in a cargo containers.
The problems that we need to solve in this paper are：
·Find the list of the number of Dones and the number of medical packages in the DroneGo disaster response system.
·Determine the number and location of containers so that drones can perform medical supply and video reconnaissance functions.
·Design specific flight routes and timelines for Drone medical supplies and video reconnaissance.
I.Assumptions
We will use the following assumptions for simplicity：
1.We regard the number of five hospitals’ emergency medical packages as a set of medical packages. Because we only know the number of medical packages required by five hospitals per day.
2.We solve the number of medical packaging into the Drone from the perspective of volume，regardless of the specific placement problem. We do this in order to simplify our model.
3.The DroneGo disaster response system does not include the H-Tethered Drone. We do this in order to simplify our model.
4.We use the same Drone to transport medical packages every day. We do this in order to simplify our model.
5.We select the locations of three containers at the ports. Because hurricane broke the road，containers can not enter.
6.The minimum altitude of the Drone in the air is higher than the altitude of all the peaks in Puerto Rico. We do this in order to simplify our model.
7.The drone's rise and fall times are equal，and the loading time and unloading time are equal. We do this in order to simplify our model.
I.1.1 Notations
II.Disaster response system model
According to us assumptions one，we can know the number of a set of medical packages is thirteen. Following that，we can get all models how to match seven Drones for meeting the requirement of transporting the thirteen medical packages. Then we establish a linear programming model to solve how How to configure the ISO cargo container.
II.1 Recommended drones and medical packages
From Drone’s max payload capability，volume and the quantity of medical packages，the constraints are
By using MATLAB，we get the programs that seven Drones to transport three medical packages.
Note that others are in tha appendix.
We choose to have the Drone load as many medical packages as possible to get the number of Drone’s model. The answer is
II.2 Cargo container configuration
We place three Drones in three cargo containers for making Drone arrive in five hospitals. Besides，we should minimize any need for buffer materials for unused space.
III Cargo container location model
We know Fajardo，San Juan and Arecibo are ports. Since road destruction，cargo container only arrives at the port. By latitude and longitude
we can calculate the distance between five hospitals finding that San Juan is the shortest distance from San Pablo and Bayamon from Table4. Therefore its Drone must have the ability of medical supply and video reconnaissance. The remaining Drones just need video reconnaissance.
From Table 4 five and six lines，we can understand that San Juan is more close than other hospitals to Bayamon and Arecibo.
III.1 Cargo container configuration in Arecibo
Let D-type Drone places in Arecibo，we can get
III.4 The result of cargo container configuration
By using MATLAB，we get following data.
From Table 4 we clearly know the ports where three cargo containers arrive at and each of cargo contains configuration.
IV.Drone scheduling and detection model
IV.1 Drone configuration and route
According to the max payload capability of E-type Drone，we draw four flight routes. Note that the number represents the flight order.
Figure 1.. Route 1
In route one，E-type Drone transports two MED1 and one MED3 to San Pablo. Next it transports two MED1，one MED2 and two MED3 to Bayamon.
Figure 2. Route 2
In route two，E-type Drone transports three MED1，one MED2 and two MED3 in all. Among them two MED1 and one MED3 belong to San Pablo. The remaining belong to Bayamon.
Figure 3. Route 3
In route three，E-type Drone first transports t two MED1，one MED2 and two MED3 to Bayamon. Next it transports two MED1 and one MED3 to San Pablo.
Figure 4. Route 4
In route four，E-type Drone transports three MED1，one MED2 and two MED3 to Bayamon. Next it transports one MED1 to San Pablo.
IV.2 Drone schedule
Since drone must be on the ground to offload cargo. So，we should calculate Drone’s fall time. Define the following variables：
is Drone charging time； 
is Drone fall time or rise time；
is unloading time or loading time. Then we can draw the total time and distance of four routes.
From Table 8，we get the conclusion that Route three is optimal. So，we get following schedule.
Where：
is the beginning of Drone take-off.
IV.3 Drone detection plan
We choose sixty cities across the major highways and roads by map of Puerto Rico. When Drones fly along these cities，they can provide road detection videos for HELP，Inc.，giving people the opportunity to provide adequate and timely response during or after natural disasters. The location of sixty cities is shown below.
Figure 5．Three video reconnaissance road maps
By Floyd algorithm，we draw three Drones video reconnaissance circuit diagram. When Drones fly from the starting point along their respective routes，the maximum range of road conditions can be detected.
Route one starts with Fajardo，and passes through the Grande，Maunabo，Yabucoa，Humacao，Panta，Santiago，Saba，then it returns to the starting point with a total distance of 73 km；Route two starts from San Juan and goes up to Vega Baja and down to Guayama. At last，it returns to the starting point along the traffic artery and passes through the densely populated areas. The total distance is 165 km；Route three takes Arecibo as the starting point，and passes through Hatillo，Saint Isabel，and finally returns to the starting point with a total distance of 230 km.
IV.4 Error Analysis
1.By use the Floyd algorithm for Drone function selection verification.，we know San Juan nearest to San Pablo and Bayamon in three ports. Therefore we decide Drone of San Juan to supply medical packages. The result matches the actual.
2.According to the configuration of the three containers，we can calculate that the medical packages of hospital Arecibo can be used 3834 days，and for San Juan is 486 days，and for Fajardo is 2273 days. These days can meet the requirement of Puerto Rico.
V. Conclusion
V.1 Strengths
1.We use the pruning method to calculate the total number of Drones in the DroneGo disaster response system，which makes the model solve simple.
2.The integer programming model provides the maximum number of medical packages per day，making it the longest to meet the needs of hospitals in Puerto Rico.
3.The model fully restores the disaster scenario in Puerto Rico with strong availability.
4.We regard the medical packages needed for five hospitals a day as a set of medical package，making our model more simplified.
V.2 Weakness
1.In the model，we assume that the packing problem is only related to the volume. We do not consider the limitation of the length，width and height of the Drone in actual packing process，which causes a certain inconsistency with the actual situation.
2.We do not consider the charging time of the drone，which may deviate from the actual transportation time.
3.A large number of assumptions reduce the accuracy of the model’s calculation results.
Memo to the manager
TO：the HELP，Inc. CEO
SUBJECT：Drone disaster relief and road detection system
Model results and conclusions
The hurricane is one of the most serious natural disasters，causing serious disasters to humanity. When a hurricane occurs，people urgently need medical supplies，and outsiders need to understand the traffic conditions in the disaster area in order to arrange rescue in time.
We develop an aerial disaster response system The drone fleet consists of one Drone D and two Drones E and a set of medical packages included seven MED1，two MED2，four MED3. The first cargo container loaded with a D-type Drone and 3834 MED1 and 1 MED2 docks at Arecibo；the second cargo container loaded with a Drone E and 2435 MED1，974 MED2，1458 MED3 docks at San Juan；the third cargo container loaded with a drone E and 2273 MED1，and 2274 MED3 docks at Fajardo. The Drone E of San Juan performs medical supply delivery and video reconnaissance，and two remaining Drones are able to detect road. We get a optimal route as follow：
Figure 1. The optimal transportation route
According this route，we calculate the shortest flight time. We draw the figure of three video reconnaissance road maps by MATLAB with Floyd algorithm. When Drones fly from the starting point along their respective routes，the maximum range of road conditions can be detected.
Suggestions
1. The Drone can carry some equipment，such as a life detector to detect the specific location of the affected person，and the local people urgently need to contact the outside world，so it can also provide wireless communication equipment.
2. Develop the function of Drones for long flight
3. The system can meet the needs of different affected areas，so we suggest that some areas susceptible to hurricanes should have our disaster response system in advance to prepare for the disaster.
4. Everyone should help the affected areas more.
Reference：
[1]Duan G B.Comparison of China，Japan and the United States disaster relief system--take Wenchuan earthquake，East Coast earthquake，hurricane Katrina as an example[J].Journal Of China Youth University For Political Sciences，2011，30（06）：115-118.
[2] Luzan B. Douglas. The enlightenment of hurricane Katrina on American disaster relief[J].China Disaster Reduction，2009（12）：14-15.
[3] Chen S X.Hurricane Katrina V.S. China’s Wenchuan earthquake--view the government s public crisis management capability from government disaster relief operations[J]. Intelligence，2009（02）：265-266.
[4] Zhang W P，Yang L Z.Practice and thinking of the army participating in disaster relief--take the US military's participation in the "Katrina" hurricane disaster relief as an example[J].China Disaster Reduction，2006（03）：38-39
[5] Zhang Q.Application of clustering pruning algorithm in outlier detection[J].Guangdong Communication Technology，2018，38（12）：58-61+75.
[6]Han Z M.Knowing the latitude and longitude to calculate the exact distance between two points[J].Technology Communication，2011（11）：196+174.
[7]Chen H，Shen J F，Zhao C.Optimization application of drones in Disaster relief[J/OL]. Electronic Technology，2019（09）：1-7[2019-01-27].http：//kns.cnki.net/kcms/detail/61.1291.TN.20181219.2134.046.html.
[8] Ding T Y，Zhang M，Fang S L.An outlier detection algorithm based on similarity pruning[J].Small Microcomputer System，2018，39（08）：1680-1684.
[9] Yue Z J，Qi H S，Xue Z.“Optimized use of drones in disaster relief” review[J].Practice And Understanding Of Mathematics，2018，48（15）：94-97.
[10] Zhang J N. Algorithm design based on linear programming[D].University Of Electronic Science And Technology Of China，2018.
[11] Peng F.Research on the efficiency of integer programming algorithm[D].Central SouthUniversity，2010.
[12]Wan Y H，Xu C F.Multi-source shortest path algorithm design for large-scale graph based on parallel computing[J].Technology Square，2017，04：68-72.
[13] Guo J，Wang X Y，Hua Y H. Design of drone track planning and monitoring system[J].Computer Measurement And Control，2018，2609：72-77.
[14] Li X M，Wang D B，Guo J K，et al.Drone track planning based on some biological heuristic algorithm[J].Machinery And Electronics，2018，3611：15-19.
[15] Wang J，Zhou S D，Zhu G T，et al.Common algorithm for drone track planning[J].Firepower And Command，2012，3708：5-8.
[16] Li B，Shi B X，Shen B. Availability constraint resource reservation and allocation algorithm[J].Computer Science，2005，02：28-30.
[17]Zhang J M，Tao Z L，Wu W J.Neural network solution for combinatorial optimization problems--the solution of the packing problem and the backpack problem[J].Journal Of East China Normal University（Natural Science Edition），1998，04：102-105.
[18]Chen D L，Chen Z Y.Model of 3D packing problem and improved genetic algorithm[J].Practice And Understanding Of Mathematics，2010，4002：142-147.
论文作者:Ma Yu-xue,Wu Zhi-feng,Yang X
论文发表刊物:《建筑模拟》2019年第10期
论文发表时间:2019/5/22
标签：kind论文; Tropical论文; Problem论文; Restatement论文; suffered论文; attack论文; storm论文; 《建筑模拟》2019年第10期论文;