Jupiter Wiki

Theme

Running the Model ​

Once you have drawn the net, specified parameters, and declared your metric queries, the next step is launching the stochastic simulation engine to produce numerical estimates.

Launching a Simulation ​

There are two quick ways to trigger a run:

  1. From the Editor: At the bottom of the right inspector panel, click Simulate.
  2. From the Analysis Screen: On the results or scenarios view, click the Simulate button on the top toolbar to re-run without returning to the canvas.

Clicking Simulate opens the configuration dialog. Once settings are confirmed, click Simulate at the bottom of the dialog to start computation.

What Happens During Execution ​

When launched, Jupiter executes an automated sequence of validation and computation steps:

  1. Static Net Validation: Checks for duplicate node identifiers, cyclic parameter definitions, division by zero, or unsupported distribution types.
  2. Expression Compilation: The internal expression compiler translates metric formulas and guards into optimized bytecode.
  3. Rust Engine Execution: The high-performance calculation engine (written in Rust) runs discrete-event simulation replications using stochastic transition matrices.
  4. Statistical Accumulation: Accumulates token sojourn times, computes point estimates, and derives statistical confidence intervals.

Monitoring and Execution Control ​

While computing:

  • A progress indicator displays completed replications and elapsed time.
  • If a long-running execution was triggered by accident (such as an excessive simulation horizon), click Cancel at any time to abort computation cleanly.

Upon completion, Jupiter automatically transitions to the Results tab on the analysis screen.

The Rapid Iteration Workflow ​

The recommended engineering approach relies on fast feedback loops:

[ Adjust Parameter / Net ] ──▶ [ Simulate ] ──▶ [ Inspect Results ]
           ▲                                            │
           └────────────────────────────────────────────┘
  1. Start with an initial simplified net.
  2. Run a fast, short simulation to verify that orders of magnitude align with expectations.
  3. Check for model inspector warnings.
  4. Increase simulation time to narrow confidence intervals.
  5. Proceed to parametric sensitivity studies in the Scenarios view.

Next steps ​

Learn how to interpret confidence intervals and model warnings in Interpreting Results.