2019年美赛优秀论文 An optional Design of DroneGo

杨利霞

2019年美赛优秀论文 An optional Design of DroneGo

Hurricane Maria Damage left Puerto Rico people a lot of pain and devastation. HELP,Inc., an NGO,
is attempting to improve its response capabilities by designing a transportable disaster response system
called "DroneGo".
In order to better assess the transport capability of the drone flfleet, we establish a Transport Capability
Evaluation Model of the drone flfleet, so that we can maximize the medicine transportation capability of
the drone flfleet when the disaster location is unknown. In this model, we use the algorithm of linear
programming to solve the limitations of volume, weight and other simple factors. After that, we used
an algorithm based on three-space segmentation and improved Monte Carlo simulation to solve the size
limitation of the objects. As a result, we obtain a reliable evaluation of the transport capability of drones,
and it can be found that among the drones, the transportation capability of drone G is the strongest.
Through the research on the reconnaissance method of the drones, we get the Reconnaissance Ca
pability Model of the drone flfleet. In this model, we fifind that the reconnaissance capability of drone
B is the strongest. Based on these two models (TCM & RCM) and abundant geographic information,
we developed a Geospatial Analysis Model. We employ the Analytical Hierarchy Process (AHP) in the
space of all position, obtaining a spatial distribution of drone flfleet ˛a′rs transportation capability and re
connaissance capability. Then, we establish an overall Effificiency Evaluation Model for the drone flfleet,
using a nonlinear programming with a variable parameter to obtain the optimal solution of the drone
flfleet overall effificiency and its spatial location. Consequently, we can determine a confifiguration of the
drone flfleet and each container’s drop point. DroneGo users can also adjust the variable parameter de
pending on whether they care more about their flfleet ˛a′rs transport capability or reconnaissance capability
to determine different options. For the problem of multiple containers, we use a dynamic programming
rule that allows users to arrange multiple containers.
In considering that the transportation capability of the drone flfleet is as important as the reconnais
sance capability, we get results that the medicine delivery can be maintained for more than two months
and road reconnaissance coverage can reach nearly 60%. Meanwhile we construct an effificient Schedule
Arrangement Model and Route Arrangement Model of drones using an integer programming model
and an improved greedy algorithm, so that the drone flfleet can be auto arranged.
The sensitivity analysis shows the strong robustness of our model. Meanwhile, we further discuss
the possibility of developing a DroneGo system software, and provide practicable advice to the HELP,
Inc. CEO.
Keywords: Multi-Objective Programming,Dynamic Programming, Drone Rescue Model, AHP