decisionanalysis

Decision analysis is a formal approach to making choices under uncertainty. It provides a structured framework for describing a problem, listing actions, specifying uncertainties, and expressing preferences. The objective is to select actions that maximize a chosen criterion, usually expected value or expected utility.

A decision-analysis problem identifies actions, states of nature, and the consequences of each action under each state. Probabilities are attached to states and numerical outcomes are assigned to action-state pairs, enabling quantitative comparison.

Common tools include decision trees, influence diagrams, and utility theory. Decision trees map actions and chance events in a branching structure; influence diagrams show dependencies among decisions, uncertainties, and objectives. Utility theory models risk preferences and computes expected utility; Bayesian decision theory integrates learning over time. Multi-criteria decision analysis addresses multiple criteria.

Typical workflow: define the problem and preferences; build the model; estimate probabilities and utilities; perform calculations, sensitivity analysis, and scenario analysis; and select an action. Implementation is followed by monitoring and updating the model as new information arrives.

Applications span business strategy, project evaluation, finance, healthcare, and public policy. Limitations include reliance on data quality and subjective judgments, potential model misspecification, and computational complexity for large problems.