Epistemic Uncertainties in Decision Making

Authors

Michael H. Faber and Marc A. Maes

Abstract

The present paper reviews and outlines the interpretation of uncertainties with a view to the various different categorizations introduced in the literature. A framework is then presented for risk based decision making taking basis in the Bayesian decision theory and recent methodical developments in risk assessment. It is emphasized that in principle all types of uncertainties should be included in formal decision analysis and that not doing so corresponds to informal decision analysis the quality of which may be difficult to judge. The controversial problem in engineering decision making of how to take into account uncertainties associated with the definition of the system being analyzed is outlined. For the typical situation where a discrete set of possible system representations is possible it is shown how a decision problem may be formulated for the identification of the optimal system to be considered as basis for decision making. The presented decision framework takes into account all prevailing uncertainties, epistemic as well as aleatory. Examples related to structural design and assessment problems relevant for offshore engineering are given illustrating how not to account for all types of uncertainties leads to sub-optimal decision making.

Published in/by

Proceedings OMAE2005, 24th International Conference on Offshore Mechanics and Arctic Engineering, Halkidiki, Greece, June 12-17, 2005, [OMAE2005-67241].