SYST 680 PRINCIPLES OF C3I

1
SYST 680 - PRINCIPLES OF C3I
Overview of Models, Gaming, and Simulation

• Dr. Rob Alexander
• robert.s.alexander_at_saic.com

2
Topics

• What is a combat simulation?
• How should they be used?
• Where is the field headed?

3
What is a combat simulation?

• Definitions, taxonomies, structure, and
  components of a force-on-force combat simulation,
  and an example.

4
Some definitions

• Model an abstract representation of reality
• Physical (iconic, e.g., windtunnel and
  airframe, map)
• Mathematical (e.g., LP, Simultaneous Eqns,
  DiffEQ)
• Logical (e.g., computer code)
• Simulation
• A model that uses numerical evaluation to imitate
  a systems operations or characteristics, often
  over time (Kelton, Sadowski, Sadowski)
• A model that represents activities and
  interactions over timeA simulation is an
  operating representation of selected features of
  real-world or hypothetical events and processes.
  (DoD)
• Wargame a simulation which depends primarily on
  humans for simulation control, decisionmaking,
  and assessment of results (may use computer for
  bookkeeping)

5
TAXONOMIES

• Static vs Dynamic
• Static no consideration of time
• Dynamic evaluate a model over time
• (DoDs definition implies that all simulations
  are dynamic)
• Continuous vs Discrete
• Continuous state of model can change
  continuously over time
• Discrete state of model can change only at
  distinct points in time.

6
TAXONOMIES

• By degree of abstraction
• Field Experiment or Live-Fire Test
• FTX (Field Training Exercise)
• CPX (Command Post Exercise)
• Wargame
• Computer Simulation
• Mathematical Model
• By means of representing reality
• Live (actual equipment)
• Virtual (crews in simulators)
• Constructive (computer simulations)

7
Taxonomies

• By degree of aggregation 3 sets of terms

ARMY
AIR FORCE

• Engineering Models
• e.g., munition effects
• High-Resolution Models
• Co, Bn, Bde operations
• Weapon versus weapon
• Medium-Resolution Models
• (not a common term)
• Bde, Div, Corps operations
• Unit versus Unit
• Low-Resolution Models
• Corps or Theater operations
• Force versus Force

• Engineering Models
• e.g., munition effects
• Milliseconds to seconds
• Engagement-level
• Aircraft and ADA interactions
• Missile versus A/C
• Seconds to minutes
• Mission-level
• Several Aircraft versus several ADA sites
• Takeoff through mission actions to landing
• Campaign-Level
• Air Campaign
• Multiple days of operations

TRAINING COMMUNITY

• Entity-Based
• Aggregate

8
Taxonomies

• By function
• Ground Combat
• Air Combat
• Naval Operations
• Log/personnel
• Air-Land
• Joint
• Operations Other Than War
• (others)

• By Durability
• Ad Hoc
• Standing
• By probabilistic nature
• Stochastic
• Direct computation
• Monte Carlo
• Deterministic
• Expected-value model (not
  mathematically correct term)
• No randomness represented

9
Some Well-Known Combat Simulations
10
Basic Elements of a Combat Model

• Preprocessor (instantiate the conceptual model)
• Tool(s) to set up input data, build a scenario,
  and test for completeness and correctness
• Simulation (move the model forward through time)
• Controls (start, pause, checkpoint, etc.)
• Simulation Infrastructure
• Mechanism to advance time (usually an event
  queue)
• Mechanism to compute state-changes for each event
• Algorithms that define state changes
• Postprocessor (analyze the results)

11
Basic Algorithms and Models of a Combat Simulation

• Entity / Unit State how the basic actors are
  defined
• Movement how the actors change location
• Target Acquisition how observers use sensors to
  acquire targets
• Attrition lethality of weapons and vulnerability
  of targets
• Decision-making autonomous and semi-autonomous
  behavior, reactive and preplanned
• Communications how messages are sent
• Reporting how information is relayed from one
  actor to another
• Situational Awareness how each actor perceives
  the state of the world

12
Eagle Combat Model An MS Case Study
13
Eagle - What Is It?

• Corps-level Combat Model.
• Force-on-Force, Unit-Based, Battalion Resolution.
• Deterministic, Time-Stepped Model.
• Balanced Across Battlefield Functions.
• Command and Control Represented Explicitly.

14
Why Was It Developed?

• Represent command and control explicitly.
• Integrate emerging (circa mid-80s) artificial
  intelligence and computer science methods.
• Rule-based reasoning.
• Automated route planning.
• Object-oriented programming.
• Reduce entire study life-cycle time, including
  scenario development and post-processing.

15
Design Philosophy

• Build on accepted combat modeling methods, but
  use AI and advanced computer science where
  appropriate.
• Use rapid prototyping software development
  paradigm.
• Use object-oriented programming paradigm.
• Develop in-house using DoD civilians and
  active-duty military.
• Carefully craft knowledge representation schemes.
• Develop balanced representation of BOSs.

16
Eagle What Does It Look Like?
17
A Brief History of Eagle
Proof-of-Concept (LANL)

DMSO C2 Exp. (HLA)
Prototype (TRAC LANL)
Prototype Postprocessor
JPSD
Trailblazer Pegasus (HLA) (JFCOM UV00)
Division Eagle
Eagle / BDS-D
Prairie Warrior
retired
Corps Eagle
Requirements (TRAC)
JVB MATREX (ACS FCS)
Small Studies
Partial VV
Planning Preprocessor
JVL
DLRC
CAA Pre- and Postprocessors
DMIF
1989
1990
Transition to AWARS
VAA (First Large Study)
1991
1988
1987
1992
1993
1994
2005
1995
1996
2003
Development activities
1997
1998
1999
2002
2000
Use in exercises or studies
18
How Did Eagles Role Develop?

• As a test-bed for distributed simulation
  technology (JPSD, BDS-D, DIS, HLA)
• As an exercise driver for staff training (Prairie
  Warrior, DLRC)
• As an analysis tool (USA Concepts Analysis
  Agency)
• As an analysis federate (Pegasus / JVB / MATREX
  for JFCOM / ACS / FCS)

19
What's In It?

• Command and Control
• Movement
• Direct Fire
• Indirect Fire
• Engineers
• Atk Helicopters
• Logistics

• Intel Fusion
• Fixed Wing CAS/Interdiction/SEAD
• ADA
• Sensors
• Commo (limited)
• HLA Interface
• C2 Interface (MCS, GCCS, others)

20
What's Not In It?

• EW
• Light Infantry
• Abn/Airmobile Opns
• NBC
• Allocation of Fixed Wing Sorties
• Weather
• Detailed communicationLOS, nets, radios

• Automated planning
• Internal staff actions of a headquarters/command
  post
• Not a functional area model for any one
  functional area
• Partial VV only (attrition, sensing)

21
How should force-on-force simulations be used?

• Purposes within DoD
• Verification, Validation Accreditation
• Principles for analysis
• Example

22
PURPOSES OF COMBAT MODELS

• Analysis
• Research and Development theoretical
  investigation of combat phenomena and design of
  weapons
• Requirements Analysis aid in understanding what
  the military needs in equipment, force
  capability, and organization.
• Test and Evaluation augments live tests in the
  materiel acquisition process
• Production and Logistics aids for studying
  industrial base, producibility, and logistics
  problems.
• Operations aids for planning on-going operations
• Training aids for preparing soldiers and units
  for combat

23
The Scientific Method

• The scientific method
• Process
• Propose a hypothesis
• Test the hypothesis
• Accept, refine, or reject the hypothesis
• Repeat
• Characteristics
• Objective
• Conforms to standard of scientific proof
• Open to examination
• Subject to peer review
• Verifiable (falsifiable)
• Combat Models partially conform to the scientific
  method.
• They provide objective, reproducible tools for
  analysis
• But they are not strictly falsifiable
• The underlying process cannot be tested
• Therefore validation is not really possible

24
V V A, plus C

• Verification is the conceptual model correctly
  implemented?
• Validation does the conceptual model adequately
  represent the relevant elements of the real
  world?
• Accreditation is the implementation approved
  for a particular use?
• Certification is the input data approved for
  use by a particular model for a particular
  purpose?

25
V V A
Specific Use (e.g., a study)
Relevant Aspects of the Real World
Purposes (Study Questions)
Modeling
Conceptual Model
Validation
Accreditation
SW Engineering (SW Rqmts)
Simulation (or other tool)
Verification
26
Principles of Good Analysis

• REMAIN OBJECTIVE!
• ANSWER THE QUESTION (Dont change the question to
  fit your model)
• Dont ask questions of a model it was NOT
  DESIGNED TO ANSWER
• Use model within design range.
• Dont limit analysis to just one model.
• Consider back-of-the-envelop analysis

27
Principles of Good Analysis

• KNOW WHY the model gave the answers it did
  (Because the model says so is never the right
  response)
• PARAMETERIZE against uncertainty (i.e.., do
  sensitivity analysis concerning subjective inputs
  and assumptions)
• Brief the ANALYSIS, not the MODEL. (If possible,
  dont even mention the model by name)

28
Principles of Good Analysis

• State your ASSUMPTIONS (Non-modelers will
  probably ascribe more power and depth to a combat
  model than it actually deserves.)
• Dont claim to provide POINT PREDICTIONS of the
  outcome of battle
• Combat models are not precise enough to do this
• Human factors play too large a part in combat
• Combat is a dynamical system subject to chaos
• The simulation may also be subject to chaos
• Instead report ranges of outcomes, trends, and
  insights.

29
Basic Principles

• Remember the scope and resolution
• Model primary process mechanisms (explicitly)
• Model secondary process effects only (implicitly)

NOTE dont expect more resolution of a model
than it was designed for. (Or, dont measure with
a micrometer and cut with a hacksaw.)
30
Basic Principles

• Ensure that each sub-model has
• a cause
• an effect (either the cause of another sub-model
  or a logged MOE)
• a model mechanism to realize the effect given the
  cause.

31
Basic Principles

• In a Force-on-Force study, evaluate each system
  in the context of its marginal contribution to
  overall force effectiveness.
• Test each system model by itself (or with other
  tightly coupled systems) by considering measures
  of performance.
• But measure the effect of each system in the
  context of all other candidate systems by looking
  at force-level measures of effectiveness.

32
MODES OF COMBAT MODEL USE

• Descriptive describe the nature of combat gain
  insights and describe trends
• Prescriptive prescribe a (perhaps optimal)
  solution to a problem
• Predictive predict the outcome of battle

33
Example Simulation as a tool in an analysis

• Future Force Warriors Exploratory Analysis
  process

FFW was an Army ATD run by Natick Soldier RDEC
from 2002 through 2007 that investigated various
individual Soldier technologies in a platoon
context.
34
Simulation for Estimating Contributions to Combat
Effectiveness

• Design a run matrix that prescribes runs using
  various combinations of the capabilities under
  consideration
• Run multiple replications of each run (for
  stochastic sim)
• Do regression analysis on the results
• A capabilitys regression coefficient represents
  its marginal contribution to overall combat
  effectiveness

35
Methodology

• EXPERIMENTAL DESIGN Vary the mix of future
  systems in each run experimental design
  specifies which systems to represent in each run.
• COMBAT MODEL simulation measures force
  effectiveness for each replication.
• SYSTEM EFFECTIVENESS ESTIMATION Fit a Response
  Surface to the results of the combat model
  Slopes of the surface estimate each
  capabilitys marginal contribution to force
  effectiveness.

Specifies Xij (presence of capability i in run j)
Computes Yj (realization of MOE for run j)
Solves for ?i (contribution of capability i) such
that ?i ??i2) (or ?i ?i ) is minimized in Yj
?j ?i Xij ?i
36
Run Matrix Example

37
Where is the field going the future of MS
within DoD

• Where has it been?
• Where is it going?

38
Where has it been?

• Past MS Initiatives
• Early Simulations
• Army Hierarchy of Models
• Army Model Improvement Program
• DMSO, AMSO, etc.
• C2 in Simulations
• Plethora of Ad Hoc Models
• Large, Monolithic Simulations
• Primacy of data

39
Where is it going?

• Distributed Simulation
• Composable Simulations
• Non-linearity Issues
• Terrain Issues
• Agent-based Simulation
• Multi-Trajectory Simulation
• Much-faster-than-real-time simulations
• Lessons from the hobby gaming world

40
More Definitions

• Distributed Simulation
• Simulators, simulations, and real systems
  collaboratively representing combat operations.
• Distributed Interactive Simulation
• Definition a set of protocols specifying how
  simulations interact on a network
• Purpose define an infrastructure for linking
  simulations of various types at multiple
  locations to create realistic, complex, virtual
  worlds for the simulation of highly interactive
  activities, running at real-time (wall-clock
  time).
• High Level Architecture
• Concept A common simulation architecture to
  facilitate the interoperability of all types of
  models and simulations among themselves and with
  C4I systems
• Definition Major functional elements,
  interfaces, and design rules, applicable to all
  DoD simulations, and providing a common framework
  within which specific system architectures can be
  defined.

41
The Distributed Simulation Vision disparate
simulations participating in a unified exercise.
42
The High LevelArchitecture (HLA)

• Architecture calls for a federation of
  simulations
• Architecture specifies
• Ten Rules which define relationships among
  federation components
• An Object Model Template which specifies the
  formin which simulation elements are described
• An Interface Specification which describes the
  way simulations interact during operation

Live Participants
Interfaces toLive Players
Support Utilities
Simulations
Interface
Runtime Infrastructure (RTI)
Federation Management Declaration
Management Object Management
Ownership Management Time Management
Data Distribution Management
The HLA is not the RTI the HLA says there will
be an RTI that meets HLA requirements but it
doesnt specify a particular software
implementation
43
Issues in Modeling and Simulation

• Interoperable and Composable Simulations -
• Impedance Mismatch, i.e., disparity of fidelity
• Algorithmic inconsistency
• Violation of the Anti-Black Box principle of
  the use of models
• Non-Linear Nature of Combat Simulations
• Small changes in input results in large,
  unpredictable changes in output
• Mathematical chaos characterizes both combat, and
  simulations, even in simple models.
• Thus, predicting the outcome of battle is
  impossible. Much analysis ignores this.
• Multi-Resolution Modeling
• Linking aggregate-level and entity-level models
• Mechanics are possible, though tedious
• Validation poses substantial problems.
• Modeling Support and Stability Opns (SASO)
• Peacekeeping and Humanitarian Relief Operations
• Mobilization and Deployment
• Anticipating hotspots in the world

44
Issues in Modeling and Simulation

• Rapid Terrain Generation
• It now takes enormous effort and funding to
  gather terrain data and create special-purpose
  terrain databases for use in simulations.
• In the age of computer-supported forces which are
  deployable anywhere in the world, terrain
  databases must be available to take advantage of
  automation.
• Terrain Visualization
• 3D real-time visualization is important for
  interactive system-level simulations
• Dynamic Terrain
• Automatic Scenario Generation
• Largely a problem of automated planning
• Standards
• For distributed simulation (DIS and HLA)
• For scenario generation (OneSAF SISO effort -
  MSDL)
• For terrain translation (SEDRIS)

45
Issues in Modeling and Simulation

• Agent-Based Simulation
• Concept autonomous agents with simple
  behavioral rules that execute those rules in an
  environment
• We hope to find emergent behavior that can help
  us understand the value of the agents rules.
• Much-Faster-Than-Real-Time Simulations
• Many applications require thousands of runs of
  slow simulations
• High-Performance Computing
• Lessons from Hobby Gaming
• First-person shooter games
• Massively Multiplayer Online Games
• Example Americas Army
• Naval Postgraduate Schools MOVES Institute does
  research in this area (among others)
• Multi-Trajectory Simulation explore alternate
  futures

46
Multi-trajectory Simulation

• Example from DARPA Course-of-Action Analysis
  Program (1997-9)

47
Summary

• Simulations are widely used as analysis,
  training, and testing tools throughout DoD.
• Use simulations appropriately follow principles
  of good simulation-based analysis.
• The MS field has many compelling challenges to
  support the military of the future.

48
Suggested Reading and Links

• Virtual Combat, David Neyland
• Introduction to Military Training Simulations A
  Guide for Discrete Event Simulationists Ernest
  H. Page and Roger Smith, Proceedings of the 1998
  Winter Simulation Conference.
• DMSO home page http//www.dmso.mil/
• The Stochastic Versus Deterministic Argument for
  Combat Simulations Tales of When the Average
  Wont Do, Thomas W. Lucas, Military Operations
  Research, Vol 5, Number 3, 2000.
• Land Warfare and Complexity, Part I
  Mathematical Background and Technical
  Sourcebook, Andy Ilachinski, Center for Naval
  Analyses, July 1996 (available for download at
  http//www.cna.org/isaac/on-line-papers.htm)
• Land Warfare and Complexity, Part II An
  Assessment of the Applicability of Nonlinear
  Dynamics and Complex Systems Theory to the Study
  of Land Warfare, Andy Ilachinski, Center for
  Naval Analyses, July 1996 (available for download
  at http//www.cna.org/isaac/on-line-papers.htm)
• http//www.ecst.csuchico.edu/hla/ (HLA training
  course notes)
• Course notes from SYST-683/OR-649 taught each
  Spring semester http//classweb.gmu.edu/ralexan3/
  SYST683