Tutorial: Create a model
4.6 • 5.0 • 5.1 • 5.2 • 5.3 • 5.4 • 6.0 • 6.1 • 6.2 • 6.3 • 6.4
This chapter shows you how to:
- Create a new model
- Create new variable nodes and enter attributes
- Move and delete variable nodes
- Draw influence arrows to define dependencies between variables
- Define variables as explicit values, functions and lists
- Invoke a built-in function to define a variable
- Do a simple Parametric Analysis
This chapter shows you how to create a new Analytica model.
In the process of building a model to analyze the costs of owning and operating an automobile, you will create variables, define dependencies, add documentary text, and compute results.
If Analytica is not open, start it by double-clicking its icon. If Analytica is already open with an active model go to the File menu, select New Model.
Creating a new model
When you start up a new Analytica session, it shows the Intro screen. Click on Blank Model to start a new model.
The blank model initially shows an Object window for an untitled model. Behind the Object window you will see an empty Diagram window.
First you should enter a Title for your model:
- Click in the Title attribute field (to the right of Title). Type in Car cost model and press Alt- enter. In the space just above the Title you entered, you will see that the Identifier field automatically changes to Car_cost_model.
- Press Tab to go to the Description attribute and type in: A demonstration model to show how to calculate the cost of running a car
Editing a diagram
In most of the three previous chapters you were in the browse mode, with the browse tool highlighted in the tool palette. In browse mode you can view an existing model without changing its structure. When you create a new model the edit tool is selected by default. You use the edit tool to create or change a model.
Be sure to note which tool is selected throughout the remainder of this tutorial.
When the edit tool is selected, a menu of icons is displayed in the node palette. These icons represent the different node types and allow you to add nodes to the diagram.
Creating variable nodes
In this chapter you will create variables in the Car Cost model.
Each variable is represented in the influence diagram as a shape that depicts the Class of Variable. These shapes are generally referred to as nodes. Select the node shape based on what you know about the variable. If you are not sure what kind of variable to use, choose a General variable .
See Tutorial: Recognizing influence diagrams for a description of the node shapes.
A new Variable Node appears in the diagram at the location where you release the mouse button.
Press Alt+Enter or click any other location in the diagram to accept the title.
Repeat Steps 1 and 2 four times to create four more variables that affect fuel cost. Title these variables as follows:
- Fuel Price (price per gallon of gasoline)
- Annual miles (number of miles driven each year)
- Mpg (miles per gallon of gasoline)
- Age (driver’s age)
Title text wraps to fit within the node as you type. You can manually create a new line in a title by pressing the Enter key at the desired break point(s). You can also expand the size of the node by dragging one of the corner handles.
Type Price and press Alt+Enter to indicate that you are finished.
Title the variables Annual Miles, Mpg, and Age.
Saving your model
Analytica automatically saves each change you make to a backup file. If your computer (or Analytica) should terminate unexpectedly, it will offer to recover your changes the next time you start Analytica. Even so, it’s a good idea to save your changes periodically.
- Select Save from the File menu (or press Control-s). The first time you try to save a model, it shows a Save As dialog, offering to save the model in the My Documents folder using the model Title as the default file name. You can select another folder and modify the file name.
- Click Save or press Enter to save the model in the selected folder.
Deleting a variable
Sometimes you might want to delete a variable that you previously created.
In this example you realize that the driver’s age is not relevant to your understanding of the Fuel Cost variable. Therefore, you will delete the Age variable.
Handles surround the node to indicate that it is selected.
Analytica asks you to confirm that you want to delete.
When you create a model you should try to structure the layout to make the influence diagram easy to understand. As you refine your model, however, you will undoubtedly want to group nodes in different ways to achieve this goal.
In this case we anticipate that Fuel Cost will be derived from the other three variables. Therefore we want to place it apart.
Moving nodes is a simple matter of dragging them to the desired position. Try moving the nodes into the new configuration below.
Editing variable titles
The name displayed inside each variable node is referred to as a Title. In this example you will expand the title Mpg to Miles per gallon:
Pause briefly between the click to select the node and the click to select the text within it. If you complete two single-clicks too quickly, Analytica interprets your actions as a double-click and opens an Object window.
If you accidentally open the Object window, return to the Diagram window by clicking the Diagram button .
The new title is displayed.
The new title is displayed.
When you change the title of a node, Analytica asks you if you want the Identifier to be automatically changed as well.
You can change this behavior, to either turn off automatic updating of the identifier or to make it fully automatic, so that you are not asked. See Preferences dialog in the Analytica User Guide for details.
Drawing arrows between nodes
One of Analytica’s most powerful features is its ability to show relationships between variables in the influence diagram. Influence arrows are used to specify the dependencies between variables.
Because the Miles per gallon variable influences the Fuel Cost variable, you will draw an arrow connecting the two nodes.
If the nodes are not connected by an arrow, repeat Steps #1 through #3.
Deleting an arrow
Occasionally, you might need to delete an arrow because of an earlier mistake or a change in your understanding of the model. This section shows you how to delete the arrow that connects Miles per gallon to Fuel Cost.
First, make sure you have the edit tool selected.
Connecting multiple arrows
When one variable is influenced by several other variables, you can draw multiple arrows at once. This example shows you how to connect the three variables contributing to the Fuel Cost variable.
First make sure the edit tool is selected.
(Hold down Control or Shift while selecting or drag diagonally across the group.)
As your diagrams become more complex, you may sometimes need to bend your arrows around nodes for clarity. Bent arrows can be either curved or crisp. Bent arrows contain way points, which define where the arrow changes directions. . You can insert or remove way points to alter the number of segments in an arrow, and drag the way points to route the arrow as you like.
When you first select an arrow without a bend, two handles appear near the midpoint -- one is a circle, the other is a square. Drag the circle to create a curved arrow initially, or drag the square to create a crisp bend. As you drag it, you introduce the first way point.
And we have two arrows with crisp bends:
To remove the bends, right-click on the arrow and select Remove all way points.
Entering attributes into the Object window
Each variable (or other Object) has an Object window that lets you see and edit its Attributes - including its Identifier, Title, Units, Description and Definition.
In this section we will use the Object window to enter attributes for the Annual Miles variable.
You can also open a variable’s Object window by double-clicking the node using the browse or arrow tools. If you are using the browse tool, you will not be able to enter or change documentation.
The identifier in the Object window shown above is Annual_miles. Analytica assigns the identifier when the title is created. It uses the first 20 characters of the title except for spaces or punctuation, which are replaced by underscores ( _ ). Analytica does not differentiate between uppercase and lowercase letters in identifiers.
You can directly edit both the identifier and the title.
First, you will change the variable’s identifier to a short abbreviation so that it can easily be used later in the definitions of other variables. You will then document the variable more fully.
Analytica uses the information from the Units field to label graphs or tables that use the Miles per year variable; Analytica does not use it in any mathematical computations.
When you change the title of a node, Analytica asks you if you want it to automatically change the identifier to match the new title.
Defining a variable as an explicit value
Analytica uses a wide range of variable types. In this section we simply enter an explicit value for the variable. Functional expressions and lists are described later in this chapter. Other variable types are addressed in later chapters.
For variables as functional expressions see Defining a variable as a function of other variables.
For variables as lists see Defining a variable as a list.
For variable as tables see Tutorial: Arrays.
For a demonstration of built-in functions see Defining a variable using a built-in function.
First let’s define Miles per year as 12,000.
Numerical suffixes like μ and K are used extensively throughout Analytica. A quick reference for these suffixes is given in Number formats.
Whenever another variable’s definition includes the identifier (Mpy) of Miles per year, this defined value, 12K, is used as its value.
Miles per year is no longer filled with a diagonal line pattern around its title, as shown in the figure below. The clear node indicates that Miles per year now has a valid definition.
Defining a variable as a function of other variables
When one variable is dependent upon another variable, you must provide an expression that describes the relationship between the variables. Influence arrows connecting other nodes to Fuel Cost show the direction of this dependency.
In this section, you will enter a definition for Fuel Cost in terms of the values of Miles per year, Miles per gallon, and Fuel Price.
Because Fuel Cost is equal to Fuel Price times miles driven, divided by miles per gallon, you will enter the following expression into the Definition field:
An asterisk (
*) represents multiplication; a slash (
/) represents division.
Based on the definition you just entered, the value of Fuel Cost is calculated by multiplying the values of Fuel Price and Miles per year, and then dividing the result by the value of Miles per gallon.
Note that Fuel Cost is no longer diagonally shaded, indicating that it has a valid definition.
Entering attributes using the Attribute panel
Sometimes you may find it more convenient to view or edit attributes of a variable in the Attribute panel as part of the Diagram view instead of using a separate Object window.
In this section, you will enter data for the Fuel Price variable in the Attribute panel.
The Attribute panel appears below the diagram. You can use this view to enter or edit data for the currently selected variable in the influence diagram.
Now that you have entered attributes for the Fuel Price variable, you will enter its definition in the Attribute panel.
In this example we will assume that Fuel Price has a fixed value of $3 per gallon.
Select Definition from the Attribute dropdown menu and enter 3.
Defining a variable as a list
In this section you will enter attributes for the Miles per gallon variable and define it as a sequence of numbers. You will use this sequence later to perform a Parametric analysis to show how the Fuel Cost is affected by Miles per gallon.
Using the Attribute panel, enter Average number of miles per gallon for the description and miles/gallon for the units.
Now that you have entered attributes for the Miles per gallon variable, you will define it as a sequence of integers between 20 and 50, by increments of 10.
Select Definition from the Attribute dropdown menu. You will notice a second popup menu appears to the right. This is the Expression popup menu. Open the Expression popup menu and select list.
Although the auto-fill feature makes it convenient to enter a simple linear sequence, you are free to edit the values as you please.
Now you will use the Expression view to see the variable’s definition in a different form. The Expression view shows the full syntax of the definition regardless of the type of variable.
Viewing results in the Result window
Now that you have entered attributes and definitions for all variables it is time to see the result. The Results button calculates values for all selected variables.
Select the Fuel Cost node and click the Results button .
The Result window appears. Icons in the upper left corner of the Result window control the view mode. Tabular form is the default.
The default view for the Results window can be changed. See the Preferences dialog in the Analytica User Guide for details.
Click the graph button to view the results in graph view.
Defining a variable using a built-in function
Analytica offers a wide array of built-in functions to simplify the process of defining variables. In this example we will take advantage of the Sequence function to re-define the Miles per gallon variable.
Let’s suppose you want to change the sequence such that the interval between Miles per gallon values is 5 instead of 10.
Select the Miles per gallon variable. Make sure that the Attribute window is open and Definition is selected.
Next, open the Expression popup menu and select Other at the bottom of the list. This opens the Object finder where you will find a large collection of built-in functions. Analytica will prompt to confirm that you want to replace the current definition of the variable. Click OK.
The Sequence function is contained in the Array library, under Creating lists. Choose Creating Lists'" under the Array library from the Library popup menu.
Scroll down the list and select Sequence.
Enter 20 for the start value, 50 for the end value, and 5 for the step. (Leave the strict field blank.) Click OK.
The new definition is displayed in the Attribute panel.
Select the Fuel Cost variable and click the Results button . The result shows improved precision with the smaller Miles per gallon intervals.
As you type textual expressions into definitions, Expression Assist provides continual context sensitive identifier completion and help on functions and their parameters. Next you will erase and re-enter the definition for miles per gallons, typing it textually this time rather than using the Object Finder, in order to acquaint you with Expression Assist.
On the diagram, select Miles per gallon and press the Object Window button on the tool bar.
Now you are ready to begin typing the definition anew. Press the ’S’ character and notice the identifier completion popup showing a list of identifiers starting with ’S’. Continue with the ’e’ character, and identifiers beginning with "Se" display:
Now, to save typing, press down-arrow three times to select Sequence and press Tab. The function name is inserted into the definition, saving you keystrokes. Now a popup shows you the parameters of the Sequence function and its description. The parameter that you are currently typing appears in bold, with optional parameter in italics.
Saving your model
After you have created part or all of a model, you should save it. Select Save from the File menu (or press control-s). Because you previously saved your model, it is saved with the original name. You can quit Analytica by choosing exit from the File menu or closing the Analytica window.
In this chapter you have:
- Started a new model
- Created new variables
- Entered attributes for variables in two different ways: using the Objects window and the Attributes window
- Re-arranged variable nodes to improve the appearance of your influence diagram
- Drawn arrows to establish relationships between variables
- Defined variables as explicit values, functions and lists
- Used a built-in function to define a variable
- Displayed the result of a simple parametric analysis
In the next chapter, you will extend the Car Cost model to include tables of data.
- Create a model
- Create and edit a model
- Create and save a model
- Create and edit nodes
- To open or exit a model
- Combining models into an integrated model
- Test and debug your model
- Tutorial: Sharing a model with ACP
- Example Models
- Example Models and Libraries
- Analytica User FAQs/Modeling issues
- Working with Large Models
- Object Window
- Definition menu
- Local Values
- Tutorial: Defining a variable as a table
- Manage attributes
- To edit an attribute
- Attribute panel
- Draw arrows
- Tutorial: Drawing arrows between variables
- Draw arrows between modules
- Parametric analysis
- Result window
- Graph view of a result
- Table view of a result
- Expression Assist
- The Expression popup menu
- Tutorial videos
- Working with Large Models
Enable comment auto-refresher