Managing Uncertainty in Systems with a Valuation Approach for Strategic Changeability

Complex engineering systems are frequently exposed to large amounts of uncertainty, as many exogenous, uncontrollable conditions change over time and can affect the performance and value delivery of a system. Engineering practice has begun to address that need, with recent methods frequently targeting such techniques as uncertainty quantification or robust engineering as key goals. While passive robustness is beneficial for these long-lived systems, designing for passive robustness frequently results in sub-optimal point designs, as optimization is forgone in favor of safety. Changeability offers an alternative means for supporting value throughout a system’s lifecycle by allowing the system to change in response to, or in anticipation of, the resolution of uncertainty, potentially enabling the system to perform ator near-optimally in a wide range of contexts. The state of the practice for valuing changeability in engineering systems relies mostly on options theory, which is associated with a number of assumptions that are frequently inappropriate when applied to change options embedded in systems. This has played a part in limiting the inclusion of changeability in fielded systems, as the standard techniques for calculating the benefits of change are often inapplicable and thus are less trusted than valuations of passive robustness. Without the ability to properly and believably value changeability, system designers will continue to look elsewhere for protection against uncertainty. A more generally applicable method for valuing changeability would greatly enhance the understanding and appeal of changeability early in the design process, and allow for the justification of its associated costs. This research has resulted in a new five-step approach, called the Valuation Approach for Strategic Changeability (VASC). VASC was designed to capture the multi-dimensional value of changeability while limiting the number of necessary assumptions by building off of previous research on Epoch-Era Analysis. A suite of new metrics (including Effective Fuzzy Pareto Trace, Fuzzy Pareto Number, and Fuzzy Pareto Shift), capturing different types of valuable changeability information, are included in the approach, which is capable of delivering insight both with and without the computational burden of simulation. The application of VASC to three space system case studies demonstrates the large range of insight about the usage and value of changeability able to be extracted with the approach. Discussion about the strengths and weaknesses of VASC is included, particularly in comparison with Real Options Analysis, and a number of promising avenues for future improvements and extensions to VASC are identified. Thesis Supervisor: Donna H. Rhodes Title: Principal Research Scientist and Senior Lecturer, Engineering Systems Co-Advisor and Thesis Reader: Adam M. Ross Title: Research Scientist, Engineering Systems