Node Definition | Data
The Node Definition dialog Data tab allows you to add or edit probability and value expressions for a chance node; and value expressions for a decision node, controlled node, or value node.
Data Input Tree
The main portion of the Data tab is the graphic representation of the Data Input Tree. The Data Input Tree allows you to enter values associated with each state of the event. For chance nodes, DPL also allows you to enter a discrete probabilities or probability distributions.
Separate Probability and Value Data Input Trees
If you have checked Separate probability and value data input trees in Node Definition | General, DPL will replace the Data tab of the Node Definition dialog with two tabs, one labeled Values and the other labeled Probabilities. Each tab has the same layout as the Data tab.
For each highlighted branch, type an expression and press the Enter key. DPL will automatically move the highlight either to the next branch (or from the probability section to the value section in chance nodes). As you continue entering data, DPL takes you through all the entries required by the node and returns you to the beginning. You can enter either numerical constants or expressions as data for either probabilities or values.
Expressions can use any of DPL's functions and the names of previously defined variables. For a list of DPL functions, click the function icon . For a list of variables, click the variable icon . Using the variable list to enter variables helps avoid spelling mistakes.
If no data is entered, DPL assumes the value for that state to be zero. (This is not true for value nodes with no data that are linked to a spreadsheet, in which case DPL assumes them to be metric nodes. See Spreadsheet Linking Overview.)
For chance nodes, DPL also allows you to enter a probability distribution. Probabilities can be defined as expressions or as Named Distributions. If you have checked Separate probability and value data input trees on the General tab or if any of the arcs leading into the chance event affect probabilities but not values (green arrowhead), or values but not probabilities (blue arrowhead), the chance node will have different Data Input Trees (and therefore separate tabs) for probabilities and values.
To reduce the number of entries you must type for a discrete chance node, DPL can automatically calculate the final value in each probability distribution once you have entered the others (by subtracting the sum of the preceding probabilities from 1.0). You may leave this entry blank if you wish DPL to calculate the remaining value. You must leave this entry blank if any probability is specified as a variable expression, otherwise you will get an "Invalid Initializer" message when you run the model.
For array value nodes, you define the number of rows and/or columns using the spin controls at the top of the dialog. These multidimensional value nodes employed to define get/pay expressions for multiple attribute models. For more on this see, Multiple Attributes and Objective Functions.
For single dimension array value nodes, you can specify that you would like to define a series by checking the Series interval entry check box. Use a series definition to define such things as a time series where the current element depends on the previous element. Intervals can be defined with starting and ending indices e.g., (From 0 to 0, From 1 to 10). Intervals must be contiguous. The $ operator allows you to use relative indices. E.g., if you wanted to define a value node called Sales that represented a ten period sales forecast with growth of 5% per period and initial sales of 1000, you could do as follows:
From To Expression 0 0 1000 1 9 Sales[$-1]*1.05
For a single dimension array value, you can specify that you would like to define a column vector by checking the Create M x 1 Column Vector check box. Elements of the column vector are denoted , , , etc.
DPL will not check any data for correctness or validity until you run an analysis.
Conditioning: Accesses the Conditioning dialog, which allows you to define the conditioning for this node.
Copy: Copies the visible portion of the Data Input Tree to the clipboard.
Full Screen: Enlarges the view of the Data Input Tree by going to Full Screen mode. This is a toggle; clicking on Full Screen again return you to normal dialog mode.
Wheel: Accesses the Probability Wheel. (Discrete Chance Nodes only)
Reorder: Allows you to temporarily reorder the conditioning nodes in the Data Input Tree.
Multiple Experts... (Enterprise and Portfolio versions): This button launches the Combine Expert Opinions dialog within which you can combine the probability assessments of multiple experts into a single likelihood or probability of occurrence. This command is only available for discrete, 2-outcome chance nodes with no probability data specified.
Zoom Commands: You can zoom the Data Input Tree in much the same way as you can in the Model Window. The four zoom buttons are Zoom In, Zoom Out, Zoom Full, and Zoom Previous. You can also zoom by selecting via Ctrl+Shift. For more on these commands see Zooming.
Font: If you'd like to change the font of the Data Input Tree you can click the Font button which launches the Font dialog. Within the dialog you can change the font type, style and/or size.
Versions: DPL Professional, DPL Enterprise, DPL Portfolio
Node Definition | Tree Instance
Reordering the Data Input Tree
Adjusting Data to Changes in Nodes or Arcs