SAE 599 Modeling and Simulation for Systems Architecting and Engineering

1
SAE 599 - Modeling and Simulation for Systems
Architecting and Engineering

• Dr. Raymond Madachy
• October 3, 2007

2
Outline

• Course project reminders
• Revisit continuous systems modeling with
  demonstrations
• Homework 2 review
• Comparison of simulation approaches
• Cargo plane solutions and tradeoffs
• Introduction to model verification and validation
• Brookss Law model testing demonstration
• Special topic transforming static models into
  dynamic simulation models
• Homework assignment
• Continuous systems modeling problems
• Stella v. 7.02 is available
• Readings

3
Course Project Reminders

• Initial status presentations next week on October
  10
• 20 minutes each
• DEN student availability
• Turn in outstanding project descriptions ASAP

4
Initial Project Status Briefings

• Plan for about 20 minutes presentation
• Describe problem and background
• what will result from your study
• Show system boundary
• How will a user interact?
• inputs and outputs
• a prototype of I/O would be very useful to show
• Reference behaviors and other methods of
  verification
• Other material as appropriate

5
Continuous Systems Examples

• Hiring delay
• Learning curves
• Exhaustion
• Rayleigh effort curve
• ODC COQUALMO defect dynamics
• Earned value
• Integrated project model
• Inspection model
• Some are discrete systems modeled as continuous

See last weeks slides
6
Outline

• Course project reminders
• Revisit continuous systems modeling with
  demonstrations
• Homework 2 review
• Comparison of simulation approaches
• Cargo plane solutions and tradeoffs
• Introduction to model verification and validation
• Brookss Law model testing demonstration
• Special topic transforming static models into
  dynamic simulation models
• Homework assignment
• Continuous systems modeling problems
• Stella v. 7.02 is available
• Readings

7
EZSIM Solutions
8
Extend Solutions
9
Extend Solutions (cont.)
10
Programmed Solution

• See http//csse.usc.edu/madachy/php/cargoplanes.p
  hp

Adding option for time reporting by events or
continuous time
11
Event Initializations
Capacity attribute attached to each plane
12
Time Loop and Event Processing
Need to know capacity of current plane in service
FIFO
13
System Dynamics Solution
Logic defect
14
System Dynamics Solution (cont.)
Logic defect
15
System Dynamics Solution (cont.)
Logic defect
16
Homework Lessons Learned

• Appropriate technique depends on goal of study
• For analysis of queues, waiting times, etc.
• Discrete event modeling is naturally suited
  because individual entities can carry attributes
• Continuous systems modeling with system dynamics
  requires excessive workarounds or non-standard
  features
• E.g. alleviate homogeneous entities with no
  individuality and no FIFOs
• Programmed solutions feasible for select problems
• Beware of effort required and resulting lack of
  generality
• EZSIM cannot easily model shared resources
• Extend can model shared resources
• Programmed solutions may be necessary when other
  tools lack specific features

17
Outline

• Course project reminders
• Revisit continuous systems modeling with
  demonstrations
• Homework 2 review
• Comparison of simulation approaches
• Cargo plane solutions and tradeoffs
• Introduction to model verification and validation
• Brookss Law model testing demonstration
• Special topic transforming static models into
  dynamic simulation models
• Homework assignment
• Continuous systems modeling problems
• Stella v. 7.02 is available
• Readings

18
Model Validation

• There is nothing in either the physical or social
  sciences about which we have perfect information.
  We can never prove that a model is an exact
  representation of "reality". Conversely, among
  those things of which we are aware, there is
  nothing of which we know absolutely nothing. So
  we always deal with information which is of
  intermediate quality -- it is better than nothing
  and short of perfection. Models are then to be
  judged, not on an absolute scale that condemns
  them for failure to be perfect, but on a relative
  scale that approves them if they succeed in
  clarifying our knowledge and our insights into
  systems.
• Jay Forrester, 1968
• All models are wrong, but some are useful.
• George Box, 1979
• Thus model validity is a relative matter. The
  usefulness of a mathematical simulation model
  should be judged in comparison with the mental
  image or other abstract model which would be used
  instead.

19
Model Validation Tests- Suitability for Purpose
20
Model Validation Tests- Consistency with Reality
21
Model Validation Tests- Utility and Effectiveness
of a Suitable Model
22
Brookss Law Testing and Improvements

• Extreme conditions for communication overhead

23
Outline

• Course project reminders
• Revisit continuous systems modeling with
  demonstrations
• Homework 2 review
• Comparison of simulation approaches
• Cargo plane solutions and tradeoffs
• Introduction to model verification and validation
• Brookss Law model testing demonstration
• Special topic transforming static models into
  dynamic simulation models
• Homework assignment
• Continuous systems modeling problems
• Stella v. 7.02 is available
• Readings

24
Outline

• Course project reminders
• Revisit continuous systems modeling with
  demonstrations
• Homework 2 review
• Comparison of simulation approaches
• Cargo plane solutions and tradeoffs
• Introduction to model verification and validation
• Brookss Law model testing demonstration
• Special topic transforming static models into
  dynamic simulation models
• Homework assignment
• Continuous systems modeling problems
• Stella v. 7.02 is available
• Readings

25
Homework

• To be assigned