Managing Multiple Sources of Epistemic Uncertainty in Engineering Decision Making

Managing uncertainty is an integral part of making well-informed engineering decisions. When formulating a design problem, many of the variables and models contain epistemic uncertainty, uncertainty due to lack of knowledge. If this lack of knowledge is significant, it may be advantageous to acquire additional information before making a design decision. In this paper, we develop a framework for identifying which sources of epistemic uncertainty should be reduced to improve the overall quality of the design decision. Using principles of information economics, utility theory, and probability bounds analysis, we determine how much additional information should be acquired for each uncertain quantity in the decision problem. Our approach is illustrated with an example for the environmentally benign design of an electric vehicle.