Vulnerability Analysis of High Dimensional Complex Systems

Abstract

We developed a neuromorphic information processing pipeline that can characterize the vulnerability of complex systems. The process consists of extracting a dynamic measure of network activity and processing the resulting time series to find patterns of collective behavior. The process succeeded in identifying extreme events that are distinct from high demand but otherwise effective system activity. Novel spatiotemporal analysis and dimensionality reduction techniques made this result possible. The pipeline can be used quite generally for analyzing high-dimensional time series and isolating extreme events in real world communication, transportation and economic systems. This system can be combined with real-time system monitoring of data streams to identify dysfunctional behaviors and characterize vulnerabilities or system instabilities as they occur.

Open PDF

Document Details

Document Type
Technical Report
Publication Date
Apr 30, 2010
Accession Number
ADA535621

Entities

People

  • Blake Stacey
  • Dion Harmon
  • Joseph Ornstein
  • Vedant Misra
  • Yaneer Bar-yam

Organizations

  • New England Complex Systems Institute

Tags

Communities of Interest

  • Energy and Power Technologies
  • Materials and Manufacturing Processes

DTIC Thesaurus Topics

  • Algorithms
  • Complex Systems
  • Computer Programs
  • Data Sets
  • Department Of Defense
  • Detection
  • Dimensionality Reduction
  • Engineering
  • Event Detection
  • Failure Mode And Effect Analysis
  • Information Processing
  • Mathematics
  • Network Architecture
  • New England
  • Pattern Recognition
  • Recognition
  • Routing Protocols

Readers

  • Educational Psychology
  • Neural Network Machine Learning.
  • Systems Analysis and Design