The Tree Instance tab of the Node Definition dialog allows you to alter the instance of a node in the Decision Tree. The tab can be accessed by double-clicking on the node in the decision tree.
Display options
The Display section offers options which affect the appearance of the node instance in the Decision Tree. A checked box next to the option indicates that it is enabled.
Show Name
On by default, this shows the name text block of a node. You can hide the name by unchecking Show Name and return it to the display by checking Show Name.
Underline Name
On by default, this underlines the node name in the Decision Tree. You can remove the underline by unchecking Underline Name.
Border Around Name
When checked, this places a border around the node name in the Decision Tree. You can remove the border by unchecking Border Around Name.
Show constant probabilities
When checked, the constant probabilities assigned to chance outcomes within Node data are displayed under each branch of the node. You can remove the probabilities by unchecking Show constant probabilities.
Show state names
On by default, this shows the name of each state of a node. You can hide state names by unchecking Show state names.
Special handing of this chance (for chance nodes only)
See Don't Gamble and Always Gamble for more information on these settings.
Outcome Grouping
This setting determines whether the node instance is Symmetric, Asymmetric, or Mixed.
Symmetric: Each branch leads to the same next event. Once placed, the node will have only one endpoint that functions as the endpoint for all branches.
Asymmetric: Each branch leads to a unique next event. Each branch will have its own endpoint.
Mixed: Some subset of the states leads to the same next event, while another subset leads to a different next event. If you select this option, when you click OK or Apply, DPL displays the Select Groups dialog. To select a subset of states, first move the highlighted bar to the first state in the subset and then hold the Shift key while placing it on the last state in the subset. Alternatively, you may select any group of states by clicking on each state while pressing the Ctrl key. These branches will share a single endpoint. Once the group is selected, press the Group button. Select the second subset of states and press the Group button again. Repeat until all states have been allocated to a group, and press OK. Press RESET at any time during this process to remove all states from groups and start over.
Objective for this decision (for decision nodes only)
It is usually appropriate for all decisions in the model to maximize or minimize the objective function and, therefore, should be defined for the model within the Objective & Utility dialog. By default, Objective for this decision radio button is set to Use Objective Function setting, meaning it will use the maximize or minimize objective function setting specified within the Objective & Utility dialog for the decision.
In rare cases it may be appropriate for a single model to have a mix: some decisions that maximize and others that minimize the expected value of the objective function. The setting in the Tree Instance tab allows you to attain this by specifying whether to maximize or minimize the Objective Function at each decision. See Maximizing or Minimizing a Decision for more information on this decision node feature.
Evaluate using Objective Function: (for decision nodes only in models with multiple objective functions)
If your model has more than one objective function specified within the Objective & Utility dialog, then you'll need to tell DPL which one to use to evaluate the tree for each decision in the model. Use the drop-down box to select the appropriate objective function for the given decision.
Versions: DPL Professional, DPL Enterprise, DPL Portfolio
See Also