What's New in Stella Professional?

Stella® Professional makes it even easier to build, share and learn from your models. Here are some of the things that are new in Stella Professional.

The Stella Window

With Stella Professional, you now have more flexibility and control over appearance, using the same set of building blocks you've always used. Stella Professional supports you step-by-step while maintaining the integrity of the models you build. You'll also find powerful new ways to work with and understand your models.

New toolbars

For previous users, Stella Professional is both familiar and new.

The first thing you'll notice is a new set of toolbars.

Build Toolbar

The Build Toolbar contains the same components as previous versions, with slightly different icons. By default, it's in the same location.

Mode Toolbar

The Mode toolbar can be used to switch between Edit mode to make changes to your models and Results mode to look at model behavior. It is allows you to navigate modules, control zoom and switch between the mapping and modeling views of structure.

Run Toolbar

The Run toolbar lets you start pause and start runs. You can also slide through the current run to view results at different times.

Moving Toolbars

All toolbars are movable and can be docked on any side of the window. By default, the Build and Mode toolbars are at the top, and the Run toolbar is at the bottom.

If you grab the handle at the left end of a toolbar, you can put the toolbar in a new location along one of the other edges, or leave it floating.

Building Models

Building is done in a very similar way to previous versions. Click a building block, then click somewhere in the Stella window to drop it in.

More control over appearance

The new look is more refined, compared to the classic Stella appearance:

You now have much more control over appearance. You can set fonts, colors, and other attributes. For example, you can change the font on flows, mark a converter as Name Only, resize stocks, and use Bezier connectors:

The Properties panel

You can set properties for all model objects from the Properties panel:

Use the Properties panel to create documentation, enter equations, set styles, and adjust many other attributes of virtually all model objects or the model itself. You can set properties for multiple objects at the same time. The set of tabs available depends on what's selected.

By default, the Properties panel is docked at the right. You can un-dock it to make it a floating window.

Model hierarchy, Run Specs, Page Settings, and other properties

• Models use a hierarchical structure. You can set styles for objects at the model level and override them at the module or object level.
• "Change the model style to" changes everything that hasn't been explicitly overridden. You can also change styles and other attributes for all selected objects.
• Run Specs and Page Settings follow this hierarchy. They're set from the Properties panel. In the root page of the model, with nothing selected, you can set the Run Specs and Page Settings for the entire model. If you create new modules, they inherit those settings, but you can override them for any module. Any modules used within a module will inherit its properties, but they can be overridden, too.
• Use the Partial Simulation tab to select specific modules or sectors to include in the simulation.
• Use the Sensitivity Analysis tab to set up a sensitivity analysis. Choose Run Sensitivity from the Model menu to run it.
• Use the Model Styles tab to set default fonts, colors, and label positions for all objects that don't already have those attributes set.

More control over layout

In addition to fonts and colors, now you can set the size on stocks and converters and set specific Converters to be name-only. This, combined with bendable Bezier connecters, allows a lot more flexibility in presentation.

It's easier to lay things out with object guides. These appear as you move objects around the window, any time one object aligns with another. For example, when you move "Module 1" next to "Stock 2", a guide line appears:

Unlimited Undo and Autosave

Stella Professional keeps a running tally of your changes so if you make mistakes, or change your mind, you can undo the changes you've made. Models are autosaved to let you recover them if something goes wrong, so you never lose more than a few minutes' work.

Exploring Behavior with Stella Live

One of the most exciting new features in Stella is the ability to explore model behavior using Stella Live. In Explore mode, you can change constants in the model by dialing them up or down, and you'll immediately see the results for building blocks, graphs, and tables displayed right on the diagram.

Enter Explore mode by running the model from the Run toolbar or menu, or select Explore from the Mode toolbar. In this mode, graphs appear on all the objects that have results available. To view larger graphs, including full axis labeling, open the Results panel and click an object containing a graph. This lets you easily walk through a model with the mouse, looking at results. Open the Results panel from the Model menu, or right-click an object and choose View Results.

You can change constants by dragging the mouse around the knobs . As you spin the knobs, the model will simulate with the new input value and you'll immediately see the results for everything, right on the diagram. You can also change values in the Results panel.

To get back to Edit mode , double-click any building block in Explore mode. All appropriate tabs will appear for the selected object.

The Results panel, graphs, and tables

The Results panel shows comparative graphs on-the-fly.

You can define and place graphs and tables in the model with these added capabilities:

• You can graph any selected variable from any selected run.
• You can include comparative and non-comparative results in a single graph or table.
• You can ctrl-drag (⌘-drag on Mac) building blocks onto output devices to populate graphs and tables. You can even use this method to populate equations, summing converters, or any input variable.
• You can use the Properties panel to control scaling, layout, and appearance.

Graphs and Tables

• Tables keep settings related to display interval and reporting options separately from the model, so you set up multiple tables with different settings to show different thing.
• Both graphs and tables let you to specify variable-run combinations so you can mix and match comparative and multi-variable results.
• You can make sketchable graphs to allow user drawn input to be compared with model results.

Comparative tables and graphs now work with the Data Manager.

• Any runs held by the Data Manager are displayed in comparative graphs and tables as soon as you specify the variable. You don't need to run the model again to see the results when you change a variable.
• If you've specified that 5 runs should be saved in the Data Manager, the oldest run will fall off of comparatives when you make a sixth run.
• During a Sensitivity run, comparative graphs and tables show the results for all of the Sensitivity runs, but once you make another run, they'll revert to only showing the runs stored by the Data Manager (which may be a smaller number of runs).
• Now you can lock individual runs in graphs and tables. You can do this on the Property panel for the graph, or for all comparatives by right-clicking and accepting the list of Displayed Runs other than Next Run.

Mapping, Modeling, and Troubleshooting

Map and Model layers

The Map and Model layers from previous versions of Stella and iThink have been replaced by views in the same window:

Choose these views from the from the Mode toolbar or the bottom of the panel.

Equation and Interface Layers

• The Equation Layer has been replaced by the Equation Viewer, which you can choose from the Model menu.
• The Interface layer isn't available in Stella Professional. We'll be offering new ways of developing and deploying interfaces.

Validity Indicators

The validity mechanism evaluates equations and units for validity and dimensional consistency.

In Model view, any unfinished or invalid building blocks are marked with a small caution sign.

There's also a list of all the issues available on the Run toolbar:

Choose any one of the error messages to highlight the invalid object. This will take you directly to the equation, even when the error is in a different module.

Once all the equations are valid, the model will automatically be checked for dimensional consistency. Any issues with units will appear in the Run toolbar; choose an error to get directed to its context. When you've resolved all the issues, the validity indicators will disappear.

Language and Function Changes

• Cycle Time lets you time stamp and track material as it flows though the model, in order to precisely compute how long different tasks and activities are taking.
• Delay Converters lets you mark a converter as changing over time in response to changes in its input, as happens with SMTH1, SMTH3, DELAY, DELAY3 and related builtins. A Delay converter shows a box inside the regular converter circle, to indicate that it has some of the properties of a stock. Because it has some properties of a stock, it's possible for Delay converters to be involved in feedback loops where no explicit stock exists.
• Units are more flexible and comprehensive, and they provide better feedback on issues with dimensional consistency. Now it's possible to define multiple aliases for a unit of measure, so that you can have more flexibility in expressing compound units.
• Sensitivity now allows for a broad range of distributions in specifying input changes.
• Module files are now stored in a folder prefixed with the model name, so you can save a separate copy of a model using the modules in the file, without changing the original model.
• The Find dialog has been updated to provide much better control for searches, including finding remaining building block errors.
• The new Random Number Generator provides more robust streams of values for the different random number builtins. The new generator can even be used with high replication Monte-Carlo.
• New Builtins have been added, mostly to support Cycle Time (CTFLOW, CTMEAN, CTMIN, CTSTDDEV, CYCLETIME, THROUGHPUT), but also UNIFORM for random number generation, and GAMMALN for statistical computations.
• The Arrayed Equation Editor for non Apply-to-All arrays makes it easier to work with arrayed variables that have multiple equations. You can pop it up from the Equation tab of the Properties panel and select multiple element entries to edit together. This makes it much easier to work with arrays where a lot of the equations are the same.
• Full Unicode Support allows you to name your variables and make comments in any language, so you can make models that communicate the way you want.

See Also