U'S' Army Engineer Research and Development Center Topics in Network Science Network Science Worksho

1
U.S. Army Engineer Research and Development
Center Topics in Network ScienceNetwork
Science WorkshopUnited States Military
Academy22-24 October, 2007

• Jeffery P. Holland, Ph.D., P.E.
• Deputy Director
• U.S. Army Engineer Research and Development
  Center

2
Agenda

• Overview
• Select Projects
• Biological and Ecological Networks
• The Network Properties of Ovarian Steroidogenesis
  in a Small Fish
• Adaptive Ecological Network Dynamics
• Cognitive and Information Systems Networks
• Shaping Suicide Vehicle Born IED Route Selection
• Mobility Common Operational Picture
• Conclusions

3
Engineer Research and Development Center (ERDC)
2000 Employees
Cold Regions Research Engineering Laboratory
Research Laboratories of the Corps of Engineers
Construction Engineering Research Laboratory
Topographic Engineering Center
Headquarters Coastal Hydraulics
Laboratory Environmental Laboratory Geotechnical
Structures Laboratory Information Technology
Laboratory
Laboratories
Field Offices
4
Overview
Network Science involves the discovery of
fundamental rules and principles underlying
systems that exhibit networking behavior Dr.
Bruce West, ARL-ARO Chief Scientist,
Mathematical Information Science,
Presentation at USMA/ARI Network Science
Workshop, April, 2007.

• Developing an understanding of the interactions
  between the network of networks across
  physical, life science, social science,
  information systems, and cognitive domains is
  necessary to achieve several major DoD and Army
  initiatives.
• We are generally at the basic knowledge
  acquisition stage in developing this
  understanding.

5
Overview (2)

• Some efforts examining systems exhibiting network
  behavior
• In biological and ecological domains
• Fish hormonal systems.
• Predator prey relationships.
• In physical, social, information systems and
  cognitive domains
• Suicide vehicle born IED traffic control point
  counter insurgency civilian interactions.
• Development of the Common Operational Picture.
• Involve multidisciplinary teams.
• Objectives are discovery of driving factors,
  structure, and fundamental relationships.

6
Network Science in Biological Ecological
Networks

• Offer unique opportunities to expand our
  understanding of increasingly complex engineered
  systems.
• Offer solution techniques that enable us to
  understand the workings of complex biological
  systems.
• Biological systems provide complex, functional
  systems in which to explore issues such as the
  relationships of network architecture to
  robustness and fragility.
• Elucidating fundamental principles/mechanisms in
  these ecological networks contributes to
  understanding network properties, design, and
  adaptation.

7
The Network Properties of Ovarian Steroidogenesis
in a Small Fish

• Purpose
• Do stressor interactions with fragile points in
  the steroidogenesis network architecture lead to
  network failure?
• Product/Results
• Network architecture of steroidogenesis.
• Mechanisms controlling network .
• Impact of energetics on endocrine function.
• Identification of network points susceptible to
  chemical attack.
• Payoff
• Ability to model robustness/fragility trade offs
  in complex networks.
• Improved understanding of how complex systems
  function.

8
The Role of Architecture in Network Fragility
Endocrine disruption in fathead minnow fish ovary
model How is the gene regulatory layer of
control integrated into the steroidogenesis
network to create a robust architecture?
output
Regulatory modules
Theca cells
output
input
Granulosa cells
output
Mitochondria
Metabolic module
Ovary steroid metabolic pathway
Hypothesis Chemical interactions with fragile
points in the steroidogenesis network
architecture lead to failure and toxicity.
Gene expression response
Chemical exposure
Reverse engineering of interaction network
Refinement of network with binding interactions
Computational model of network architecture
Defining network architecture
Examine chemical effects on network architecture
9
Adaptive Ecological Network DynamicsHunter-Prey
Relationships

• Purpose
• To numerically and mechanistically describe how
  learning and environmental heterogeneity
  contribute to adaptation in hunter-prey networks.
• Product/Results
• Theory and algorithms describing predator prey
  relationships.
• Infrastructure to port work to multiple systems.
• Payoff
• Ability to predict impact of animal learning and
  communication on the information propagation
  affecting survival in hunter-prey networks.
• Hierarchical network dynamics in static versus
  dynamic heterogeneous environments.

GENERIC SYSTEM OF AGENTS WITH NEUROLOGICAL
CAPABILITIES
Observers Perspective
Animals Perspective
10
Hunter-Prey Relationships The Role of Cognition
Networks in Survival

• 3) Strategic algorithms
• complex behavior selection based on forecast of
  expected rewards

1) Movement algorithms move track animal in
physical environment
2) Tactical algorithms encode information in
neural pathway elicit response
Goodwin et al. (2006)
Carlile et al. (2006) Rind and Simmons (1999)
Anderson and Steele-Feldman (2006)
11
Network Science in Cognitive Information
Systems Networks

• Offer opportunities to expand our understanding
  of complex adaptive systems.
• Provide a venue in which to explore issues such
  as the relationships of data, information,
  knowledge, and understanding.
• Networks comprising the contemporary operational
  environment and battle command are complex,
  adaptive, and interactive.
• These include networks of networks with physical,
  social, information systems, and cognitive
  domains.
• Identifying driving factors and insights in these
  systems contributes to understanding of
  underlying relationships.

12
Shaping SVBIED Insurgent Behavior and Mission
Outcome
ERDC-USMA Team

• Purpose
• To develop methods for capturing complex adaptive
  system behavior to explore factors associated
  with Traffic Control Point (TCP) strategy
  effectiveness and SVBIED mission outcome.

Problem Suicide Vehicle Born Improvised
Explosive Devices (SVBIEDs) pose a persistent
threat, impacting unit operations, U.S. policy
and public perception.1 They are difficult to
detect or defeat as they maneuver intermixed with
locals.

• Products/Results
• Means to explore driving factors regarding
  SVBIEDs in the operational environment.
• Insights regarding TCP strategies robust against
  a range of SVBIED behaviors.

• Payoff
• Increased understanding of key factors.
• Enhanced ability to examine strategies.

1 FMI 3-34.119/MCIP 3-17.01 Improvised Explosive
Device Defeat, Sept. 1005, Chapter 2, page 1.
13
Shaping SVBIED Insurgent Behavior and Mission
Outcome
Entity Interactions - Agent Based Modeling
Global Path Planning Artificial Electromagnetic
Field Theory
Large-Scale Experimental Design Nearly
Orthogonal Latin Hypercubes

• Efficient Space filling
• Factors associated with
• TCP Network
• Situational Awareness
• SVBIED Capabilities
• Target Selection

Approach
Outcomes
Mission Outcome by SVBIED Reaction Type
Factor Correlation with Outcome
14
Mobility Common Operational Picture
ERDC-NPS-USMA Team

• Problem Our militarys success in
  network-centric operations (NCO) is threatened by
  our lack of fundamental knowledge concerning
  network interactions across physical,
  information, cognitive, and social domains.

• Purpose
• To create a baseline unified knowledge space for
  shared awareness of ground vehicle mobility and
  maneuver.
• Product/Results
• Data model and ontology.
• Limited demonstration of embedded semantic
  reasoning for tactical maneuver in dynamic
  routing.
• Payoff
• Advances in modern semantic approaches for
  military decision making.
• Contributions to the Common Operational Picture
  for mobility and maneuver in the battlespace.

15
Mobility Common Operational Picture
Conceptual demonstration architecture
Context for Common Operational Picture
OWLViz displaying the inferred hierarchy for the
class Segment
Scenario snippet with mapping to ontology,
reasoner, and services
16
Conclusions

• To understand complex systems, we must be able to
  relate interactions at the fundamental level with
  outcomes at the macro-scale.
• The way these processes interconnect (the network
  of networks) is itself an important part of the
  science.
• We are making strides and adding to the body of
  knowledge as basic understanding and insights are
  built.