Adaptive planning and uncertainty resolution

Abstract

The article discusses uncertainty resolution and the approaches to uncertainty in planning. Statically perceived uncertainty may be described more or less explicitly in decision making. Stating various possible outcomes and, possibly, their numerical probabilities may allow quantitative theory to be applied. However, there will be a residual of genuine uncertainty due to the limits set by the available information. By “buying” more information the decision maker can decrease the genuine uncertainty. It is also often possible to reduce uncertainty by waiting. The most effective way is to state the expectations explicitly and to let the various future states of information correspond to alternative decisions in a conditional strategy. This requires a new kind of forecast that refers to the development of information available. A less efficient mode of planning can still take into consideration the fact that revisions are expected. This is flexible planning in a wide sense.