High-fidelity computational thermal models (HFMs) of mechanical systems typically incorporate multi-disciplinary data sources to define boundary conditions, constraints, and dynamic system inputs. Oftentimes, HFMs are used during the planning, design, fabrication, testing, and operational phases of the mechanical systems, however, most of that data is processed during the modeling and test phases to discover and verify system responses. This approach can lead to much unused data and engineering effort that could otherwise provide useful information during the operational phases of the systems. One major bottleneck in using HFMs during the operational phase is data volume and computation time. Reduced-order models (ROMs), such as Gaussian processes, can consolidate data volume, data complexity, and time complexity needed for processing HFMs. The Borg multi-objective evolutionary algorithm (MOEA) presents a possible effective approach for processing ROM information in conjunction with real-time true process data to better understand the real-time state of a system. An investigation is being performed into the use of ROMs with the Borg MOEA to capitalize on engineering effort and simulation data that would otherwise be abandoned. This paper discusses the results of such a study in a steady-state conductive-radiative heat transfer system.