In most cases, if you change the number of states for a node, change the node's type, or change the type of influence for an arc into the node, you will have to edit the node data, including data in dependent nodes. You can access the Data tab of the Node Definition dialog by double-clicking on the node.
If you change the number of states for a discrete chance node, you will need to edit the node data, as the probabilities for the states will not add up to 1.0 anymore. Open the Data tab of the Node Definition dialog for the node and you will see the new states reflected in your Data Input Tree.
If you delete an influencing decision, discrete chance, or controlled event, or reduce the amount of influence in an arc, your Data Input Tree will be reduced to reflect this. If the node was conditioned on the influencing event, then it would have had different sets of probability (or value) data associated with each state of the influencing event. Now that the influencing event is deleted, the node needs only one set of probability (or value) data (assuming the node is not conditioned by anything else). DPL saves the set of data that is closest to the middle of the original tree and assigns it to the amended, uninfluenced (or less influenced) node. Data assigned to the other branches will be lost. (DPL will display the message "Node data will be lost" before deleting any data.)
If you add an influencing event, or increase the amount of influence in an arc, DPL will copy the original node data to the new subtree for each state of the influencing node. You may edit the data to reflect the dependencies on the influencing event.
If you add a new state to an influencing event, the input tree of the dependent node will be expanded to accommodate the new state. DPL will not assign any data to the new branches.
If you delete a state of an influencing event, all data associated with that state will be deleted, and probabilities will no longer sum to 1.0.
Versions: DPL Professional, DPL Enterprise, DPL Portfolio