Language Constructs for Identifying Flexible Points in Coercible Simulations

1
Modeling and Simulation Technology Research
Initiative
Language Constructs for Identifying Flexible
Points in Coercible Simulations
Joseph C. Carnahan Paul F. Reynolds, Jr.,
Ph.D. David C. Brogan, Ph.D.
2
Modeling and Simulation Technology Research
Initiative

• Simulation Design and Transformation for Reuse
• COERCE
• Coercibility the practices and methods for
  capturing designer knowledge in software
  (Carnahan et al.)
• Coercion a user-guided, semi-automated software
  transformation process (Waziruddin et al.)
• Composability
• Reusing components, possibly with acceptable
  amounts of revision, to meet new requirements
  (Bartholet et al.)

3
Overview of COERCE
Simulation concept
Initial requirements
Simulation instance
Coercible Simulation
Simulation Coercion
Simulation instance
Simulation instance
Flexible point metadata
Reuse requirements

4
Flexible Points

• Any element of the simulation that can be
• manipulated in meaningful and effective ways
• to direct a simulations behavior
• Examples
• Value substitution for constants/parameters
• Changing loop convergence criteria
• Adding and removing stochastic elements
• Allowing execution order to vary

5
Clarifying Flexible Points

• Pragmatic definition, emphasis on semi-automated
  transformation
• Contrast with
• Simulation parameters
• Design decisions

Parameters
Flexible points
Design decisions
6
Tools for COERCE

• Focus is on flexible points
• Language constructs (to describe)
• Optimization techniques (to manipulate)
• Visualization tools (to display and evaluate)

Work presented here on language constructs
7
COERCE and Language Constructs

• Need a way for developers to describe COERCE
  metadata
• Many flexible points evident at development time
• Desired metadata properties
• Easy to write
• Machine-readable

8
Language Construct Requirements

• Useful information about a flexible point
• Code that must be changed to manipulate it
• Alternative bindings that are available
• The original intent and implementation
• Effects of changing a flexible point
• Which simulation variables are affected and how
• What the behavioral effects (performance and
  accuracy) of each alternative are
• How the effects of different flexible points
  interact with each other

9
Complexity of the Challenge

• Meeting all of these requirements is difficult
• Compounded by the wide variety of flexible
  points
• Model versus model-implementation
• Narrow versus broad
• Ordered versus unordered alternatives
• Independent versus entangled

10
Language Construct Ideas

• Two possible approaches
• A toolbox of language primitives
• Different constructs for different types of
  flexible points?
• Different constructs for different available
  information?
• A single killer construct
• Tradeoff between flexibility and complexity?

11
Flexible Point Table

• One solution Tabular representation
• Rows for alternatives
• Columns for
• Name
• Description
• Possible implementation
• Can be implemented as an embedded XML document,
  using XSLT for presentation

12
Flexible Point Example
13
Evaluation of Table Construct

• Applicable to a significant set of flexible
  points
• Turning constants into variables
• Swapping constant parameters with stochastic
  functions
• Replacing arbitrary block of code (function body)
• Provides opportunities for automation through
  COERCE
• Easy to implement with existing languages and
  tools

14
Future Directions

• Extending the flexible point table
• Looking at flexible points as cross-cutting
  aspects
• Aspect-oriented languages and systems
• Special constructs for describing broad flexible
  points
• Identifying flexible points without location in
  code
• Additional exploration through examples

15
Summary

• Explored definitions for flexible points in
  coercible simulations.
• Investigated variety of language constructs for
  capturing coercibility information, and
  identified a set of requirements for them
• Created the flexible point table as a method for
  addressing capture of flexible points
• Evaluated utility of flexible point table

16
Acknowledgements

• Defense Modeling and Simulation Office
• National Science Foundation
• MaSTRI at the University of Virginia

Modeling and Simulation Technology Research
Initiative