Outline SAL The Systems Architecture Lab Mission Approach - - PDF document

5/22/2006

• A. H. Levis

1

5/22/2006

System Architectures Laboratory

1

SAL

Architectures, Command Centers, and Adversaries: Design and Modeling

Alexander H. Levis alevis@gmu.edu SAL

5/22/2006

2

SAL

System Architectures Laboratory

Outline

• The Systems Architecture Lab

– Mission – Approach – Research Projects

• Research Areas

– Architectures – Command Centers – Adversarial Modeling and Effects Based Operations

• Conclusion

5/22/2006

• A. H. Levis

2

5/22/2006

3

SAL

System Architectures Laboratory

Mission

• The System Architectures Laboratory conducts basic and applied

research in the modeling, analysis, design, and evaluation of – System Architectures – Organizational Architectures (coalition, adversarial) – Effects-based Operations – Information Operations

• In all cases, the emphasis is on Command and Control applications.

5/22/2006

4

SAL

System Architectures Laboratory

Approach

• The research focus is on temporal aspects and dynamics:

– Adaptive Organizations / Command Centers – Executable models of C4ISR Architectures – Timed Influence nets and Dynamic Influence nets – Planning and Execution Monitoring – Dynamic Assessment

• We use discrete event system theory in general and Colored Petri nets

in particular to develop algorithms and then embed the algorithms in tools

• We are taking a systems engineering approach; one of the objectives

is to understand systems of systems

• We are building a test bed that contains tools, data, and models and is

accessible remotely by collaborating institutions

5/22/2006

• A. H. Levis

3

5/22/2006

5

SAL

System Architectures Laboratory

Current Research Projects

• ONR: Adaptive Organization Design for Effects Based Operations
• AFOSR: Interactive Planning for Capability Driven Air and Space

Operations

• AFOSR: Computational Modeling of Adversary Attitudes and

Behaviors (with Carnegie Mellon)

• AFOSR: Human Centric Design Environments for C2 (with Berkeley

and Vanderbilt)

• AFRL/IF: Dynamic Air and Space Effects Assessment (Critical

Experiment) (DASEA)

• Raytheon: Timed Influence Nets support

5/22/2006

6

SAL

System Architectures Laboratory

Architecting…

• Alternative approaches and tools can be used to design the

Architecture views – Structured Analysis or Object Orientation

• The executable model is used for logical, behavior and performance

evaluation

• The main challenge is the development of measures and procedures

for the evaluation of architectures

Architecture Design Executable Model Construction Architecture Analysis and Evaluation M I S S I O N DoD AF Product Generation DoD Arch. Framework

The Complete Process

Operational Concept

5/22/2006

• A. H. Levis

4

5/22/2006

7

SAL

System Architectures Laboratory

On Architectures

• Current research on Architectures is focused on the development of

executable models that represent an architecture. – Both structured analysis and object oriented approaches are used. – The resulting discrete event dynamical system models are expressed as Colored Petri Nets. – Analytical, algorithmic, and simulation tools are used to analyze the behavior of the modeled architecture . and to evaluate performance.

• While there are methodologies, tools, and techniques for the design of

architectures for individual system, there is very little theory and virtually no tools for the design of Systems of Systems

• Systems of Systems are currently assembled together in an ad hoc

manner; they are not designed.

5/22/2006

8

SAL

System Architectures Laboratory

Evolution of Systems Engineering

• Since WW II, systems engineering has continued to evolve addressing

increasingly complex systems

• Since the mid 50s we have been focusing on increasingly large

integrated systems

• The challenge is now to address loosely coupled Systems of Systems

Complexity SoS SE and Enterprise SE focus for R&D

5/22/2006

• A. H. Levis

5

5/22/2006

9

SAL

System Architectures Laboratory

SoS Examples

• Air Traffic Control (NGATS)

Multiple Carriers FAA

• Allied and Coalition Operations

Multiple Airport Authorities

5/22/2006

10

SAL

System Architectures Laboratory

System of Systems*

A system will be called a System of Systems (SoS) when: – The component systems achieve well-substantiated purposes in their own right even if detached from the overall system; – The components systems are managed in large part for their

• wn purposes rather than the purposes of the whole;

– It exhibits behavior, including emergent behavior, not achievable by the component systems acting independently – Component systems, functions, and behaviors may be added or removed during its use

* Definition evolved from Sage, Maier, Levis, …

5/22/2006

• A. H. Levis

6

5/22/2006

11

SAL

System Architectures Laboratory

A Key Challenge

• In the Systems of Systems, many humans are inside the SoS
• But our methodology and procedures really consider the human as being
• utside; we design “interfaces”
• We do the Task Allocation too early along traditional lines
• The Task Allocation problem requires some new thought and algorithms;

adaptation requires the ability to change dynamically the allocation (morphing)

Hardware Humans Software Systems Engineering

Integration Human-Computer Man-Machine

5/22/2006

12

SAL

System Architectures Laboratory

Degree of Integration

Integration covers the spectrum from tight integration into a single large scale system to interoperation in a loosely coupled system of systems

Interoperating System of Systems (systems enter and leave) Current Stovepipes

Other Domains HUMINT IMINT SIGINT Intel Defense Suppression

Combat Ops Mobility

SOF

Air Defense Other Domains Other Domains HUMINT IMINT SIGINT Intel Intel Defense Suppression

Combat Ops Mobility

SOF

Air Defense Other Domains Netw ork/Information Sharing Infrastructure HUMINT IMINT SIGINT Intel Defense Suppression

Combat Ops Mobility

SOF

Netw ork/Information Sharing Infrastructure

Combat Ops Combat Ops Mobility Mobility Air Defense

Degree of Coupling

Fully Integrated (Monolithic) Uncoupled Loosely Coupled Tightly Coupled

5/22/2006

• A. H. Levis

7

5/22/2006

13

SAL

System Architectures Laboratory

Traditional Systems Engineering Approach

In a systems-centric viewpoint, functions are tightly coupled to systems

System a System b System c System f System g System h System e System d System d Function r Function S Function t Function w Function v Function x Function u Function y Function z Activity 1 Activity 2 Activity 3 Activity 4 Activity 5 Activity 6 Activity 7 Activity 8 Activity 9

Operational Activities System Functions Systems

The SE Design Problem: Map Operational Activities to Systems Functions

5/22/2006

14

SAL

System Architectures Laboratory

Service Oriented Architecture (SOA)

• In a service-oriented viewpoint, activities drive services and services drive systems

– effectively decoupling operational activities from systems

This is counter to traditional systems engineering

System a System b System c System f System g System h System e System d

Operational Activities Functions

Function r Function S Function t Function w Function v Function x Function u Function y Function z Activity 1 Activity 2 Activity 3 Activity 4 Activity 5 Activity 6

Enterprise Services

Service A Service C Service B Service D Service F Service G Service E Service H Service I Service J Service N Service M Service L Service O Service K Activity 7 Activity 8 Activity 9

Systems

Loosely Coupled

5/22/2006

• A. H. Levis

8

5/22/2006

15

SAL

System Architectures Laboratory

A Layered Architecture for Virtual Integration

• To achieve loose integration, a layered SOA architecture is needed that

includes the nine Net Centric Enterprise Services

9 Net-Centric Enterprise Services

Messaging Enterprise Systems Management User Assistant Mediation Storage Collaboration Discovery Services IA/Security Services Application

Mission 1 Mission 2

Infrastructure Domains COIs Mission

5/22/2006

16

SAL

System Architectures Laboratory

Command Centers and EBO

• This research effort focuses on several problems:

– Adaptive architecture for coalition command center organization – Integration of conventional and information operations – Course of Action development and evaluation tools – Dynamic Assessment of effects – Modeling an adversary’s attitudes and behaviors

• An organization may need to adapt because:

– The task load (adversary action) has changed – The mission has changed – The resources available have changed

• A key example is the Navy’s Expeditionary Strike Group
• Development of courses of action that integrate all elements of

national power (DIME; PMEESI) requires models that describe an adversary’s attitudes and behaviors

5/22/2006

• A. H. Levis

9

5/22/2006

17

SAL

System Architectures Laboratory

Timed Influence Nets

• Influence nets – a specialization of Bayesian nets – have been used for

more than ten years in determining adversary’s pressure points and actions that may lead to desired effects

• The development of Timed Influence Nets has added

– Temporal logic to Influence Nets – Dynamics

• The theory has been implemented in the form of a tool for the modeling,

and analysis of dynamic influence nets

• Pythia version 1.1 has been released and is being installed on the

testbed

• Applications include

– Course of Action development – Course of Action analysis and comparison – Modeling of adversaries – Computational models for culturally influenced behavior in colaitions

5/22/2006

18

SAL

System Architectures Laboratory

Pythia 1.1

• Timed Influence Net Modeling and Analysis tool
• Developed with support from ONR and AFRL (and

initially with support from AFIWC)

• Enables analysts to create executable (probabilistic)

models that link potential actions (elements of a COA) to effects based on knowledge about the environment

• Captures the rationale for COAs that explains how actions can achieve

effects – Given a set of actionable events, determine the Courses of Action that maximize the achievement of desired effects as a function of time

5/22/2006

• A. H. Levis

10

5/22/2006

19

SAL

System Architectures Laboratory

The Modeling Approach

Set of Desired Blue End States

Command Intent

Ops Pol Trans Fin Rel IO

Set of Blue’s potential Actions that will affect Red.

Time-phased broad actions Desired End States Set of Desired and Undesired Effects Probabilistic model relating actionable events to effects through a network of influencing relationships: Influence Net model From Red’s Point of View May include Red’s COAs

Effects Actions

Model Construction

Example: Suppressing IEDs

International Influences CF Actions Political and Religious Structures Economic/Infrastructure Insurgents Activities Local and Regional Support Overall Effects

5/22/2006

• A. H. Levis

11

5/22/2006

21

SAL

System Architectures Laboratory

Probability Profiles

• Assigning temporal information (time delays) to

the links and nodes in the model, converts the influence net to a timed influence net.

• One can experiment with the set of actions and

their timing and see the probability of different effects as a function of time.

• Changing actions or timing can effect the

likelihood of various effects over time.

• The COA can be “tuned” to generate the best

probability profile.

• The COA being tested shows that it will take 6

months to significantly reduce IEDs, they will be reduced more on the A route than the B – Note the cross over of local versus regional support for insurgency after about 21 months Weeks

5/22/2006

22

SAL

System Architectures Laboratory

Capabilities

• COAs can be compared
• Example shows previous COA

compared to one that using diplomatic efforts to reduce international support for insurgents for the effect that IED attacks on Route B are suppressed

• Pythia offers a set of other algorithms

and utilities to support the evaluation of the COA and the model. – Sensitivity Analysis – Set of Actions Finder – COA selection and “optimization” – Incorporation of evidence Weeks

5/22/2006

• A. H. Levis

12

5/22/2006

23

SAL

System Architectures Laboratory

Conclusion

• The Lab is addressing a number of contemporary problems

– via basic research (6.1) – via applied research (6.2, 6.3)

• It also applies the research results to current problems

– Expeditionary Strike Group – Broad Area Maritime Surveillance – CAOC

• It maintains collaborative relations with research groups in other

universities (as prime or subcontractor)

• It has close relationships with DoD organizations and AFCEA
• It has close relationships with industry (BAE Systems, Boeing,

Lockheed Martin, Raytheon)