Adaptive Decentralized Resource Optimization

Abstract

We propose a new paradigm for distributed resource optimization which is considerably more general than allpreviously proposed models and captures the complexity inherent in military operations: time-dependent objectives, local as well as global constraints, couplings between neighboring nodes. Our model is motivated by problems of positioning for optimal monitoring, fusion of disparate types of data in sensor networks, and adaptive routing in search and-rescue scenarios. We will develop algorithms for our distributed optimization model, and by extensions for all of the above problems, which are fully distributed, scale favorably in the number of nodes, and exhibit fast geometric convergence. Furthermore, we will develop techniques that will allow for switching adaptively between distributed and centralized protocols. Moreover, we will deliver codes implementing our methods that will be accompanied by transparent guarantees on performance

Document Details

Document Type
DoD Grant Award
Publication Date
Feb 03, 2017
Source ID
N000141712195

Entities

People

  • Alexander Olshevsky

Organizations

  • Boston University
  • Office of Naval Research
  • United States Navy

Tags

Fields of Study

  • Computer science

Readers

  • Adaptive Control and Estimation with Uncertainty in Dynamic Systems.
  • Computer Networking
  • Distributed Systems and Data Platform Development