Challenges of Simulating Uncertainty of Information

Abstract

Human information interaction including human computer interaction is complex given the variability of the human and now the ever-increasing amount of information available. While more information seems to be ideal, most would agree that having the needed information at the time required maybe the ideal. This is particularly true for decision-making. At each step along the decision-making process, different information may be needed so that the goal of generating the best decision is reached. However, information is not perfect and comes with uncertainties. Our approach has been to explore what we have termed uncertainty of information (UoI) and its impact on decision-making. We have selected Gershon’s taxonomy of imperfect information as variable descriptors to represent potential sources of uncertainty. Using terms from the taxonomy we have created an algorithm that can capture the uncertainty from this point of view. In addition, we have developed simulations to begin to explore the how UoI from various sources can influence decision-making. Consequently, we have been faced with the difficulties of building a general comprehensive model for simulating UoI. These difficulties range from how to express the behavior of the human to how to express the behavior of the computer and devices that may be providing the information. The difficulties also include how to incorporate the different factors and underlying dependencies that contribute to the UoI value.