Algorithmic Approach for Network Inference and Monitoring: Coding over Real Numbers

Abstract

First, we want to characterize the optimal number of measurements that are needed to identify any problematic component of a connected complex system. Second, we aim to provide constructive methods to generate those measurements. Finally, we would like to test these theoretical results by numerical simulations. All three goals have been essentially met. More concretely, for goal number 1, we have shown the lower bound of number of measurements for any given graph. In particular, for several special graphs, we were able to find the exact optimal number of measurements. Furthermore, for general networks, we were able to provide order optimal estimation of the number of measurements needed. For goal number 2, all above estimation comes with explicit measurement construction. In particular, for general networks, we have a construction which only uses polynomial time algorithms. Finally, for goal number 3, all above results are extensively tested numerically.

Open PDF

Document Details

Document Type
Technical Report
Publication Date
Mar 28, 2014
Accession Number
ADA624778

Entities

People

  • Ao K. Tang

Organizations

  • Cornell University College of Engineering

Tags

Communities of Interest

  • Energy and Power Technologies

DTIC Thesaurus Topics

  • Algorithms
  • Complex Systems
  • Compressed Sensing
  • Computer Programming
  • Construction
  • Department Of Defense
  • Detection
  • Information Theory
  • Linear Systems
  • Measurement
  • Monitoring
  • Numbers
  • Real Numbers
  • Recovery
  • Signal Processing
  • Simulations

Readers

  • Approximation Theory.
  • Computational Linguistics
  • Systems Analysis and Design

Technology Areas

  • AI & ML
  • AI & ML - Bayesian Inference
  • AI & ML - Machine Learning Algorithms