ABSTRACT
This chapter presents methodologies decision trees and influence diagrams, which allow one to map from decisions/alternatives to a probability distribution over an outcome space. Decision trees and influence diagrams have been applied to decision making in such diverse areas as portfolio management, commercialization of a new drug, and scheduling of the refueling for a nuclear power plant. The chapter also presents sensitivity analyses of decision trees and influence diagrams and discusses the use of expected utility as a performance measure within decision trees and influence diagrams. Branches that emanate from a decision node represent a set of comprehensively exhaustive, mutually exclusive decisions at a point in time. The basic objective associated with the analysis of a decision tree is the selection of the initial decision that optimizes expected monetary value for the entire time frame represented by decision tree. Decision trees are somewhat arbitrary for any particular situation since typically a decision situation may be represented by one or multiple decisions.
