Classes of decision analysis

The ultimate task of an engineer consists of developing a consistent decision procedure for the planning, design, construction and use and management of a project. Moreover, the utility over the entire lifetime of the project should be maximized, considering requirements with respect to safety of individuals and the environment as specified in regulations. Due to the fact that the information with respect to design parameters is usually incomplete or uncertain, decisions are made under uncertainty. In order to cope with this, Bayesian statistical decision theory can be used to incorporate objective as well as subjective information (e.g. engineering judgement). In this factsheet, the decision tree is presented and answers are given for questions on how new data can be combined with prior probabilities that have been assigned, and whether it is beneficial or not to collect more information before the final decision is made. Decision making based on prior analysis and posterior analysis is briefly explained. Pre-posterior analysis is considered in more detail and the Value of Information (VoI) is defined