Incorrect decision tree

[Manuel Luque]

The decision tree shown by OpenMarkov for the attached decision analysis network is incorrect. It presents two problems:

1. All the decisions should appear in the decision tree.
2. The correct maximum expected utility is 4.

I don't know exactly the connection between the decision trees evaluation and the package org.openmarkov.inference.variableelimination. If they are connected then I could help to solve this issue.

[Manuel Luque]

Please, Miguel, it is better if you wait a bit and first Íñigo and I talk to see if we left these things ordered after August changes.

[Miguel Ángel Artaso Landa]

No problem at all. Let me know whenever you want me to take over it.

[Iñigo Bermejo]

There was a bug in the code that calculated whether an ordering was relevant by checking whether the set of nodes revealed by a decision was a subset of those revealed by another. In this case both sets were empty and therefore a subset of each other.