Emerging artificial intelligence methodologies in uncertainty analysis and modeling

Abstract

This paper presents an introductory overview of various methods used in representing and solving uncertainty problems. The emerging methods help us to capture both the subjective judgment and incomplete information and data in decision analysis under uncertainty. The type of method to be used for a problem depends on the situation. Although, the emerging methods appear to address uncertainty very well, problems are usually encountered. The emerging methods involve complicated mathematics, which are fairly difficult to understand and handle for the average transportation and pavement engineer. As the sample space grows, it becomes very difficult to handle and analyze uncertainties using the emerging methods. The solution becomes difficult and labor intensive. The availability of software will be a very promising step. Some of the emerging methods like influence diagrams and valuation-based systems handle asymmetric problems.