Home / Monte Carlo / Analyzing the Model / Run Model
Run Model¶
From Excel click...
QXL Monte Carlo New > Monte Carlo > Run Model
Runs the simulation. Each trial draws a value from every marked input's distribution, recalculates the model, and collects the value of every marked output. The results go to a new report worksheet.
Run Model reads the two lists beside it on the ribbon and nothing else. The Simulations list sets how many trials, and the Sampling list sets how the input values are drawn. Both are described on the Options page.
The two evaluators¶
Quantum XL states the choice itself, in its own words:
Quantum XL evaluates a simulation one of two ways: with Excel's own calculation engine, or with the Quantum XL Banyan-Tree Math Parser, which is more than 100 times faster.
| Evaluator | What it is |
|---|---|
| Quantum XL Banyan-Tree Engine (Fast Mode) | every formula in the model was translated, so the engine evaluates the trials itself |
| Native Mode - Excel Calculated Monte Carlo | one full Excel recalculation of the workbook per trial |
You do not choose between them and the engine does not guess. Before a run is trusted, Quantum XL compares the translated model against Excel on a set of inputs. Fast Mode is used only when every formula translated and the two agreed on every value. Anything else falls back to Native Mode, and the report says which one ran.
Check before a long run. Validate Model answers the same question on its own and names the exact formulas holding a model back, without running a simulation first.
Before it runs¶
The marking checks run first, as they do for every command that reads the model. If they have something to say, they say it and the run does not start.
While it runs¶
A status window shows the stages and offers Cancel. It runs on a thread of its own, so it keeps repainting and keeps answering that button while the simulation itself is blocking.
Custom Run¶
QXL Monte Carlo New > Monte Carlo > Custom Run holds four more commands for a run that does not match the two ribbon lists:
| Command | What it does |
|---|---|
| Run Fast Mode | twelve named counts, in a Monte Carlo group and a Sobol (Quasi-Random) group, asking for Fast Mode |
| Run Native Mode | three named counts, 1,000 to 100,000, asking for Native Mode |
| Custom... | asks for any count from 10 to 100,000,000 |
| Set up Histogram Groups... | decides which outputs share an overlaid histogram in the report |
Each Custom Run item names its own sampler on the button. A count listed under a Sobol heading runs Sobol whatever the Sampling list on the ribbon happens to say, so a menu item never changes under you because of a list somewhere else.
What is not saved¶
Neither ribbon list is stored in the workbook. Both go back to their defaults every time Excel loads the add-in. The same is true of the Histogram Groups dialog and the custom count.
Related¶
- Options, for every count each list offers and both Custom Run dialogs.
- Validate Model, which reports the evaluator without running a simulation.
- Percent Contribution, which decomposes an output's variation across the inputs.