Decision Trees
Visually Map Decisions and their Possible Consequences
Gain Probabilistic Insights in Your Decision Analysis
In a complex world, the decisions we face are often more sophisticated than “pick X or Y.” Previous decisions and uncertainty shape the choices yet to come. Each specific decision path we could take leads to unique situations, yet we must make choices now prior to knowing precisely where we will find ourselves in time, down the road. Which decisions will lead to the best returns, or lowest costs?
Decision trees allow us to visually model these chronological layers of decisions and random events to determine the best business policy under uncertainty. They are well-suited in the analysis of multi-stage decisions over time, where values at each stage are uncertain.
What are Decision Trees?
Visual Decision Modeling Across Industries
Decision trees are a model type that accounts for the conditional nature of future decisions, giving realistic and useful decision modeling analytics. The technique is used in construction & engineering, energy & utilities, mining & minerals, logistics & transportation, consulting & legal, healthcare & pharmaceuticals, and many other disciplines. Use cases include:
- Resource extraction strategies
- Bidding decisions
- Build vs. buy strategies
- Litigation planning
- Treatment planning
- Negotiation strategies and more
Prescriptive Decision Suggestions
Decision trees give a decision maker an overall policy suggestion to utilize throughout the life of the project, as well as a probabilistic comparison of the risk profile of different potential sets of decisions. There is also great value in being able to visualize large, complex decisions and the various stages of decision-making, which decision trees do exceptionally well. The visual nature of decision tree diagrams makes them well-suited to discussion, problem-solving, and communication to others.
How do Decision Trees Work?
Decision trees are a structure of linked nodes, starting with an initial node (the first choice or unknown you will encounter), then branching out to all the ensuing possibilities. Node types represent decisions or random (chance) events, and the cost or return of each branch from each node is indicated. Probabilistic likelihoods for each outcome from a chance node are also estimated and included in the model.
The paths represent sequences of events, ending in the calculated value for that sequence as well as for every node in between. Each node is independent of nodes not on the path that led there, and thus the structure of decisions and events that ensue can be completely unique from any other node or path.
With the conditional value of each node now calculated, a decision tree will determine the optimal path or top decision policy for the entire network of possibilities, maximizing or minimizing the overall outcome as appropriate.
A decision tree showing decision and chance nodes, as well as payoff values and probabilities.
Decision Tree Software from Palisade
Palisade’s PrecisionTree software puts powerful decision tree analysis at the fingertips of any Excel user. PrecisionTree makes it easy to create, analyze, and share trees with others, all from the familiar Excel environment. Furthermore, PrecisionTree is designed to work with Palisade’s @RISK and TopRank products, which add Monte Carlo simulation and sensitivity analysis to decision tree models. These combined analyses allow the creation of the most accurate tree models anywhere, while maintaining point-and-click ease-of-use.