Real-world implementations of geographic resource allocation solutions

Abstract

The purpose of this project is to conduct real-world experiments that determine the efficiency of algorithms that we have developed" in previous and ongoing projects with the Office of Naval Research. In our prevous project with ONR, ~Online and decentralized algo""rithms for geographic resource allocation problems~, as well as our ongoing project, ~Online and decentralized algorithms for ~horse""fly~ problems~, we have designed numerous algorithms for solving avariety of geographic resource allocation problems with online an"d decentralized characteristics. A geographicresource allocation problem is any mathematical optimization problem in which spatial" properties such as distance, shape, connectivity, contiguity, perimeter, area, or convexity play an important role. A few simple ex"amples of problems of this kind include: ? Determining the most efficient routes for a vehicle (or a fleet of vehicles) to visit a c"ollection of targets to performsurveillance, provide service, or collect data.? Partitioning a given geographic region into smalle"r districts to provide service to a set of customers or to equitably balance an existing geographic asset.? Selecting the optimal l"ocations of service facilities to provide service to a region or a group of regions.In our previous and current projects, we have" constructed proven theoretical guarantees as well as computational simulations that show that our algorithmic solutions are of high" quality. However, we have not yet implemented these solutions on actual hardware; thus, we have not taken into account many importa"nt real-world considerations. The University of Southern California has recently leased a warehouse that serves as a shared research laboratory for autonomous aerial vehicles and ground testing that includes facilities for motion capture and trajectory analysis. I"n thisproject, we will conduct experiments in this warehouse using autonomous aerial, terrestrial, and aquatic vehicles to move our"" algorithms from the theoretical, ~proof-of-concept~ phase to real-world implementations, thus verifying that our solutions do indee"d perform well in practice.

Document Details

Document Type
DoD Grant Award
Publication Date
May 05, 2017
Source ID
N000141712487

Entities

People

  • John Gunnar Carlsson

Organizations

  • Office of Naval Research
  • United States Navy
  • University of Southern California

Tags

Fields of Study

  • Computer science

Readers

  • Canine Service Warrior Training Program for Wounded Warriors in the Veterinary Industry, Supported by Donors.
  • Distributed Systems and Data Platform Development
  • Operations Research