Date of Award

12-2025

Document Type

Dissertation

Degree Name

Doctor of Philosophy (PhD)

Department

Mechanical Engineering

Committee Chair/Advisor

Gregory Mocko

Committee Member

Gang Li

Committee Member

Satchit Ramnath

Committee Member

Laura Redmond

Committee Member

Cameron Turner

Abstract

This dissertation considers several challenges facing the systems engineering community. These challenges stem from the proliferation of digital design methodologies, where design decisions are made on the basis of simulation models, with the goal of delaying costly physical prototyping as much as possible. Simulation-based systems design enables the consideration of larger design spaces, complex technologies, multiple stakeholders, and challenging use cases and mission scenarios. Several challenges have arisen in this area of research, and this dissertation aims to address three of these challenges. The first is model fidelity evaluation to support the selection of appropriate simulation models for a given design task. Validating a model for which no physical prototype yet exists becomes a problem of mapping a first principles understanding of physics to the mathematical formalizations contained within the model. The second challenge is reconciling with both inherent, irreducible uncertainty in a system, and uncertainty stemming from a lack of information or ignorance about physical phenomena that affect system behavior. The third challenge involves the scale of systems design problems, which are often subject to the curse of dimensionality. Additionally, decisions among several criteria, with multiple stakeholders, and subject to uncertainty challenge the assumptions about perfect rationality that several decision methods rely on.

To address these challenges, three research thrusts are explored. The first is a physics-based model fidelity evaluation method, focusing on comparative, qualitative model fidelity rather than absolute fidelity quantification. The second thrust aims to evaluate the robustness of a system with a high-dimensional input space, subject to several failure modes or limiting phenomena. This method enables the construction of systems performance measures that consider both the raw performance of a system and its robustness to irreducible variance. The third thrust addresses making decisions with high-dimensional input and output spaces when the system preferences are poorly formed, or when there is an aversion to preference models requiring an early commitment to a set of preferences that may later change with additional information. These challenges are addressed together in the context of a military ground vehicle design case, where multiple inputs, outputs, preference models, and fidelity levels are considered in a design problem with uncertainty. This dissertation ultimately works towards simulation-based systems design methodologies that bridge the gap between a designer’s imperfect rationality and lack of information, and computational methods for solving large, complex design problems under uncertainty.

Author ORCID Identifier

0009-0005-8792-9852

Available for download on Thursday, December 31, 2026

Share

COinS