Electronic Thesis/Dissertation


A Bayesian Network and Real Options Methodology for Engineering Infrastructure Investment Decisions Under Uncertainty Open Access

Downloadable Content

Download PDF

This research develops a new, fully quantitative methodology to evaluate engineering infrastructure projects built under highly uncertain market conditions through a three-stage multi-factor decision model. The first stage is a Bayesian network model built to quantify each project's future cash flow in terms of key economic and financial variables and to estimate market uncertainty. The second stage uses real options analysis to calculate the value of an option to defer the time (in years) to invest in a project. The third stage uses decision analysis to identify financial alternatives that fall within the investment decision solution space. The investment decision solution space contains the set of all feasible financial values that also satisfy the constraints of the investment problem. The model is an improvement over traditional financial analysis that uses net present value (NPV) alone as the investment decision criterion. NPV is inflexible and ignores the effect of uncertainty on investment decisions, especially when financial stakes are high and market conditions are volatile. By contrast, the proposed model is flexible, using stochastic variables that exploit uncertainty by quantifying the effect of volatility on investment decisions. As a result, a decision agent avoids investing prematurely in capital-intensive real assets when market conditions are unfavorable. The option to delay an investment reduces overall corporate risk and enables an organization to redirect financial resources toward more favorable near-term investments.

Author Language Keyword Date created Type of Work Rights statement GW Unit Degree Advisor Committee Member(s) Persistent URL