Case Studies
TLDR: Case studies are runnable templates with deeper evidence attached. Read the writeup for source data, model translation, validation, and limits, or open the same model directly in the template library.
Use This Like A Buyer Or Analyst
What to look for in a professional case study
- Read the question first. A good case study makes the decision problem clear before it starts showing charts.
- Check the validation. The point is not to look plausible; it is to show what the model matches and where it still abstracts.
- Open the template only after you know the lesson. Then use the recommended first experiment to see the model behavior yourself.
Fastest Path
Want to explore instead of just read?
Every published case study has a one-click model path. Use the writeup to understand the reasoning, then open the live template when you want to inspect or test the abstraction directly.
Published Case Studies
BPI 2017 Offer Log
Public loan-offer event log translated into a staged DES model with weighted routing, calibrated cycle time, and intervention analysis.
Sepsis Cases
Loop-heavy public healthcare log showing where a lean staged DES abstraction starts to diverge from richer simulation detail.
What Makes A Strong Case Study
| Element | Why It Matters |
|---|---|
| Named data source | Shows the model is grounded in something auditable instead of invented assumptions. |
| Explicit translation choices | Helps buyers see what the model preserves and what it simplifies. |
| Validation table | Separates a plausible diagram from a defensible simulation. |
| Intervention results | Demonstrates that the model can support decisions, not just description. |
| Limitations section | Builds trust by stating what the model does not claim. |
Quick-start templates teach mechanics. Case-study templates show judgment: how to pick the right abstraction, how to calibrate it, and how to explain the result in business terms.
Use The Published Model
Both published case studies are now available as loadable templates inside Simply Solver. Use BPI when you want a cleaner staged-flow benchmark, and use Sepsis when you want to stress-test a denser, loop-heavy healthcare pathway.