A neuro -fuzzy decision support system for risk analysis of revenue -dependent infrastructure projects

Luis Henrique Martinez, Purdue University

Abstract

Risk analysis is one of the most important processes that every decision-maker has to perform before undertaking any project. It involves a deep understanding of all the factors that affect the output of the project. This is especially important in revenue-dependent infrastructure projects, where the life of the project depends almost exclusively on its revenue output. The analysis of revenue-dependent infrastructure projects requires the involvement of a large set of qualitative and quantitative factors. The sponsors need to analyze the influence of both factors before making a decision to proceed with the project. Due to the limitations of the typical risk analysis process, the qualitative factors are analyzed separately from the quantitative factors. Usually the qualitative factors are the first to be analyzed, forcing the decision-maker to provide a subjective decision before the analysis of the quantitative factors is performed. This approach could constrain or even neglect potential opportunities. This research provides a decision support system that combines the quantitative and qualitative factors for risk analysis of revenue-dependent infrastructure projects. The system considers the relationships among these factors and the influence that they can have on the development of an infrastructure project. It recommends a neuro-fuzzy technique that is able to model the relationships between the risk factors. This neuro-fuzzy model was designed using data from an actual infrastructure project. The analysis of both types of factors in a single tool will contribute to a reduction of the subjectivity involved in the risk analysis process and will provide better analysis to the decision-maker. The main advantage is that the decision-maker will be able to include in the analysis factors that do not have a numerical nature. Moreover, if the system is properly fed with historical data, it will provide an estimate of the output based on both types of risk factors, quantitative and qualitative.

Degree

Ph.D.

Advisors

Halpin, Purdue University.

Subject Area

Civil engineering

Off-Campus Purdue Users:
To access this dissertation, please log in to our
proxy server
.

Share

COinS