DecisionTool creates decision trees with reusable action spaces and chance spaces for enumerating all possible outcomes and calculating expected value of a multi-dimensional utility vector for all terminal nodes.
- Create trees composed of nodes, discrete action spaces and discrete chance spaces.
- Implement modular spaces (action and chance) so that a space can be located at many nodes, so that it represents the same decision or random event.
- Assign multi-dimension utility to any node
- Assign probability to the options of a chance space
- Calculate multi-dimension utility of any terminal node
- Calculate probability of any terminal node
- Add graphical representation of the decision tree.
- Concatenate several spaces as the representation of a cross product between their options
- Add cotinuous action spaces
- Add continuous chance spaces