Simulation and What-If Runs

Use this guide when a Physical AI scenario needs to be tested with simulation or what-if analysis before engineering review.

Run Flow

Prerequisites

Requirement	Why it matters
Review question	The engine should match the decision being reviewed.
Input owner	Simulation results depend on trusted assumptions and constraints.
Runtime access	Some engines or adapters require project configuration.

Inputs

Input	Review focus
Scenario package	Asset identity, geometry, layout, process rules, and constraints.
Engine selection	DES, EnergyPlus-oriented building simulation, Monte Carlo, system dynamics, surrogate, or what-if.
Parameters	Duration, random seed, changed variables, horizon, and confidence settings.
Output record	KPIs, run status, export files, error notes, and reviewer.

Procedure

1. State the question and expected decision.
2. Choose the engine that matches the question.
3. Prepare inputs and assumptions.
4. Run a single scenario or batch study.
5. Review output KPIs, confidence, and limitations.
6. Store run evidence with scenario ID and reviewer notes.
7. Decide whether to accept, revise, or run another study.

Expected Output

The output is a simulation review package with engine choice, input assumptions, run outputs, limitations, and decision notes.

Validation Checklist

• Engine choice matches the question.
• Parameters and changed variables are recorded.
• Output KPIs are tied to scenario and run ID.
• Limitations and assumptions are visible.
• Reviewer decision is captured.

Failure Handling

Symptom	Response
Engine does not match the question	Select a more appropriate study type before comparing results.
Run fails	Record error, input version, and runtime state before retry.
Output lacks confidence	Treat result as exploratory and gather more runs or field data.
Assumptions change	Create a new run record rather than overwriting the prior result.

Related Docs

• Physical AI Tools
• Layout and Process Planning