Conditioning

Conditioning allows you to vary the values and/or probabilities associated with a node depending on the states of other nodes.

How to Condition Nodes

Conditioning is done in one of two ways: either in the Data tab of the Node Definition dialog accessed by double-clicking on the node you would like to condition, or by double-clicking on an arc whose type you would like to change.

See Node Definition | Data and the Conditioning dialog for information on how to specify conditioning via node definition. See the Change Arc Type dialog for information on how to specify conditioning via changing an arc's type.

Note: If the conditioning in the influence diagram contains cycles, DPL will return an error message when it compiles the model.

Influence Arcs

In a DPL influence diagram, an influence arc between nodes indicates dependence, conditioning or timing. When you condition a node, DPL automatically draws the appropriate influence arcs in the influence diagram. Alternately, you may draw arrows yourself and then change their type to specify conditioning.

Colored Arrowheads

The color of an influence arc's arrowhead indicates the type of influence:

Orange - Both values and probabilities depend on the state of the influencing event

Black - Neither probability nor value data depend on the state of the influencing event; black headed influence arcs are used to indicate timing in the decision tree

Green - Probability data depends on the state of the influencing event, value data does not depend on the state of the influencing event

Blue - Value data depends on the state of the influencing event, probability data does not depend on the state of the influencing event

If you prefer to see only black arrowheads that do not indicate the type of influence, select File | Options | General and uncheck the box labeled Colored arrowheads (arcs).

Modifying Influence Arcs

If you change a node's conditioning, DPL automatically updates the relevant influence arcs in the Influence Diagram to reflect the changes. However, you may also create, delete, and modify arcs directly. For more information see Influence Arcs.

Versions: DPL Professional, DPL Enterprise, DPL Portfolio

See Also

Influence Arcs

Bayesian Revision

Node Definition | Data

Conditioning Dialog