Repository logo
Log In(current)
  1. Home
  2. Colleges & Schools
  3. Graduate School
  4. Doctoral Dissertations
  5. PERSEVERANCE: Predictive Engineering with Resilient Simulation and AI-Enabled Reasoning
Details

PERSEVERANCE: Predictive Engineering with Resilient Simulation and AI-Enabled Reasoning

Date Issued
May 2026
Author(s)
Jones, Gerald
University of Tennessee at Knoxville
Advisor(s)
Xueping Li
Additional Advisor(s)
Bing Yao
Thomas Berg
Jamie Coble
Permanent URI
https://trace.tennessee.edu/handle/20.500.14382/57526
Abstract
Modern socio-technical systems operate under increasing uncertainty and consequence, requiring decision-support methods that can represent human behavior, infrastructure constraints, and policy interactions under disruption. This dissertation presents PERSEVERANCE: Predictive Engineering with Resilient Simulation, Enhanced Reasoning, Virtual Environments, and Novel Computational Engineering, a unifying framework that integrates simulation, artificial intelligence (AI), and optimization to support resilience analysis across critical domains. The first study focuses on mission-critical asset management and maintenance planning. It introduces a suite of AI Asset Management Assistants (AMAS) that employ retrieval-augmented generation (RAG) and task-specific large language model (LLM) agents to support maintenance request intake, hazard identification, and automated work package generation. System performance is evaluated in terms of retrieval accuracy, document completeness, and planning efficiency, demonstrating that AI-assisted planning can reliably support high-consequence maintenance workflows while remaining interpretable and auditable for human planners. The second study addresses disaster response through an AI-driven flood evacuation simulation. A multi-persona archetype methodology is introduced in which LLMs generate heterogeneous, behaviorally grounded agent actions with structured explanations. Embedded within an agent-based modeling framework, this approach enables scalable population-level simulation while preserving behavioral diversity. Results illustrate how persona distributions, environmental conditions, and coordination constraints influence evacuation outcomes, positioning LLM-driven agents as exploratory behavioral models rather than prescriptive optimizers. The third study develops a hybrid system dynamics and agent-based model of the opioid epidemic. Heuristic optimization techniques are used to calibrate system dynamics parameters against historical overdose and treatment data, establishing a validated macro-scale baseline. AI-driven agents representing distinct behavioral archetypes are then integrated to explore how social determinants of health, treatment accessibility, and policy interventions shape overdose and mortality outcomes. Collectively, this work demonstrates how AI can augment simulation to improve interpretability, behavioral realism, and exploratory policy analysis within a coherent systems engineering methodology.
Subjects

Simulation

Machine Learning

AI

Work Packaging

Flood Response

OUD

Disciplines
Industrial Engineering
Public Health
Systems Engineering
Degree
Doctor of Philosophy
Major
Industrial Engineering
File(s)
Thumbnail Image
Name

JonesJrDiss_2026.pdf

Size

5.71 MB

Format

Adobe PDF

Checksum (MD5)

413727b7231b7106fce181e611919942


University Libraries

1015 Volunteer Boulevard
Knoxville, TN 37996
865-974-4351

Map & Directions
Donate to the Libraries
  • About
  • John C. Hodges Society
  • Speaking Volumes magazine
  • Outreach
  • Directory
  • Employment
  • Policies
  • Library Intranet
University of Tennessee power T logo

The University of Tennessee, Knoxville
Knoxville, Tennessee 37996
865-974-1000

Events
A-Z
Apply
Privacy
Map
Directory
Give to UT
Accessibility

Built with DSpace-CRIS software - Extension maintained and optimized by 4Science