How many endpoints (sometimes called leaf nodes) can DPL Release 6.0 handle?
There is no fixed limit. At some point, either your computer’s memory is exhaused or (more often) decision analysis runs take longer than you would like to wait. Every model is different, but most users find that with a cashflow spreadsheet of moderate size they are able to work with thousands of endpoints. For a model without a spreadsheet, or with a spreadsheet which has been converted to DPL code, millions of endpoints (or more) can be tractable. • Should I start building my DPL model with an Influence Diagram or a Decision Tree? This is a matter of taste. Most decision analysis practitioners recommend starting with an influence diagram, but for models with a conditional structure (i.e., “If we decide not to do the marketing test then …”) starting with a tree is often more intuitive. • Why can’t I run a Monte Carlo simulation on my (discrete) model? If your model is a discrete tree (i.e., it doesn’t contain any continuous chance nodes), a Decision Analysis run with the Discrete Tr
