Common Sense Is Not So Common Some Leasons Learned Using Simulations for Problem Solving

1
Common Sense Is Not So Common Some Leasons
Learned Using Simulations for Problem Solving

• Charlie Voss
• Golder Associates Inc.
• October 25, 2007

2
Topics

• GoldSim to the rescue or pass me the hammer
• Collaborative model development
• Stochastic is fantastic not
• Verification, calibration, documentation and
  other tions
• Presenting your masterpiece
• GoldSim karma

3
Before You Start

• GoldSim may not always be the best solution
• Dont force or let your clients force you into
  using GoldSim
• Save it for complex, dynamic, stochastic
  simulation problems
• Sell the approach, not the software

4
Model Development

• Top-down right way to go, often hard to
  practice
• Collaborative development
• You should have taken more psyche courses
• The quality of the model depends on the skill and
  purpose of the modelers
• Find a way to give everyone some ownership in the
  model
• Look out for the quiet guy
• Critical in a regulatory environment
• Keep an eye on the data commitment

5
Probabilistic Are you sure?

• Results are often incomprehensible to your
  audience
• Too much certainty in the uncertainty?
• Correlations often difficult to quantify
• Deterministic with sensitivity analysis may be a
  better approach

6
Verification and Documentation

• Spreadsheets are your friends
• What time is it?
• Fortran, GoldSim some things dont change
• Start explaining model results early
• Incorporate links to report sections, spreadsheet
  checks, etc.

7
Presenting Results

• GoldSim or Lunesta?
• Never underestimate your graphics department
  everyone can relate to graphical representations
• Bar charts and tables old school but cool
• Unless youre modeling a parking lot!

8
My favorite GoldSim Project
9
Giving back

• Look for opportunities to get involved with local
  schools and educators
• You may benefit more than those youre helping