Docstoc

Introduction to Wavelets and Wavelet Transforms

Document Sample
Introduction to Wavelets and Wavelet Transforms Powered By Docstoc
					Survivability of Large Scale Networks and Design Research
NSF-EXCITED Workshop February 28, 2005 Distinguished Professor of Industrial and Manufacturing Engineering The Pennsylvania State University University Park, PA 16802
skumara@psu.edu

Soundar R.T. Kumara

CYBER DESIGN NET(CD-NET)
Retrieval Agent

Retrieval Agent Design Agent 2

Design Agent 3 Design Agent 4 Coordinator Agent Design Agent 1 Design Agent n Customization Agent

Retrieval Agent

Retrieval Agent

Retrieval Agent

Company A Agent

Company B Agent Repository Agent

Design Agent A

Agent based Design Network

Design Agent B Univ. Agent

Idaho

MIT BU SU UM NIST VT NASA PSU

GT

Web service Repository / Digital Libraries

Agent based Design
Design Agent 1 Design Agent 2 Customization Agent Coordinator Agent Design Agent n
Customer Needs Functional Requirements

Retrieval Agent

Design Repository

Retrieval Agent
Design Repository

Retrieval Agent

New Product
Product Analysis

NSF-ITR : An Information Management Infrastructure for Product Family Planning and Mass Customization, PI: Timothy W. Simpson (PSU), Co-PIs: Soundar R.T. Kumara (PSU), S.B. Shooter (Bucknell), J.P. Terpenny (Virginia Tech), R.B. Stone (U. Missouri-Rolla), August 2003 – July 2006

Logistics Network
Agent Based Logistics Network

General Motors: Development of Wireless based Automatic Deployment and Load Makeup System PI: Soundar R. T. Kumara (PSU). (January 2001 – current)

Sensor Networks

NSF SST : Self-Supporting Wireless Sensor Networks for In-Process and In-Service Integrity Monitoring Using High EnergyHarvesting Nonlinear Modeling Principles. PI: Soundar R. T. Kumara (PSU) Collaborators: S. Bukkapatnam (Oklahoma State), S.G. Kim (MIT) and X. Zhang (UC Berkely) (September 2004 – August 2007); Marine Corps: Integrated Diagnostics: Soundar Kumara and Barney Grimes

Military Logistics (UltraLog)
•Secure against cyber attack •Robust against damage •Scalable to wartime data loads

UltraLog: Extremely survivable net-centric logistics information systems for the modern battlefield
DARPA - ULTRALOG : Chaos, Situation Extraction, and Control: A Novel Integrated Approach to Robust and Scalable Cognitive Agent Design PI: Soundar R. T. Kumara (PSU) (Jan. 2001 to July 2005)

UltraLog Challenges (PSU)
 Situation Identification  Performance Estimation  Adaptive Control

 Hierarchical Control
 Robustness  Infrastructure level  Application level  Network Survivability  Security

Methodologies
 Chaos based time series analysis, Machine

learning

 Digital sensors

 Model predictive control
 Auction mechanisms  Mathematical optimal control

 Queueing theory
 Complex networks theory

Situation Identification
62% 5 6 7 52% 10

100% 1
TAO

59% 8

64%

62% 11 13

4
9

95% 2

12

Objective: Estimate
64% 14 15 16 17

global stress environments at TAO

Methodologies: Time
series analysis (Chaos), Machine learning

Adaptive Control
500 100

N3

500

A5
1000

A6

A7

LP
Heuristic

N1
500

N2
A8 A4 A9 A10
100

Objective: Build
N4
A11 CPU

A1
200

A3

distributed adaptive control policy for the stress environment

A2

A12

Control facilities:
Resource allocation, Alternative algorithms

500

N5
A14 A15

300

A13

A16

Adaptive Control
Methodologies: Model predictive control, Auction

Stress Environment Continuous Modeling
Sensor Sensor Sensor

Sensor Design Mathematical Programming

Agent 1

Agent 2 Periodic Auctioning Auction

Agent 3

Decentralized Coordination

DMAS Implementation: CPE Society

 Military logistics
  

Command and Control Structure Distributed, continuous planning and execution Stressful Environment: Stresses range from heavy computational loads to infrastructure loss

 Objective: Identify and demonstrate key concepts in the

argument for and concept of “design for survivability

Specification and Performance Estimation
 Methodology: XML based

distributed specification (TechSpecs), Queueing theory based performance modeling.  Description:




TechSpecs described agent attributes, measurement points and control parameters. BCMP network and Whitt QNA employed to estimate the end-to-end app-layer response times and remove infeasible operating modes.

Control of the DMAS
 Methodology: Application-

Layer control using queueing theory, and other learned models.  Description: Trading off QOS (plan quality) for performance (response time) using estimates gained from Queueing network models. Regression models used to assess the impact of model prediction on application utility.

Designing a Network Infrastructure
 Methodology: Optimization

using GA.  Description: Represent the entire network of agents as a math programming model with constraints on resources with an objective to minimize the total set-up costs.

Hierarchical Agent Society Satisfying Constraints with Minimum Total Infrastructure Set-up Cost

Mathematical Formulation

Load Control Problem for Agent Systems
 Optimal resource control to optimize long run

performance.  Piecewise deterministic Markov process for dynamic environment (workload and CPU availability)
workload stress
r1, Z ( t )

h( x(t ))
x1 (t )

Z(t) ~ finite state, CTMC

d1
l1

B
r2, Z ( t )

CPU
DZ (t )

x 2 (t )

d2

CPU stress
d1, d2 : CPU time allocation l1, l2 : algorithm control

B Workload

l2
c2 (l(t ))

Agent

Survivability: Topological perspective
 Objective: Survivability of large-scale network  Methodology: Complex networks theory

Cyber Design Network (CD-NET)
Company A Agent Company B Agent Repository Agent

Challenges:
•Security against cyber attacks, hackers •Robustness against damage (infrastructure and application) •Scalability to growth and load of the network

Design Agent A Design Agent B Univ. Agent

Robustness

Idaho

MIT BU SU UM NIST VT NASA PSU

GT

Web service Repository / Digital Libraries

Scalability

Security

Distributed Large Scale Networks Research- Lessons Learnt and their usefulness to CD-NET
 Distributed Agents – Agent definitions,

communication and platform are critical

 Agent Composition to solve a problem is feasible

through TechSpecs (meta-data) and dynamic service discovery

 Ontologies are the foundation for TechSpecs

 Infrastructure Survivability – Optimization

approaches

 Application Survivability – Through CAS analysis