Model Fidelity in Mission Scenario Simulations for Systems of Systems : A Case Study of Maritime Search and Rescue

Systems of Systems (SoS) represent a view where individual, independent systems named Constituent System (CS) collaborate to achieve overarching capabilities not attainable by a single CS alone. In cyberphysical systems, CS can be engineered entities like boats or aircraft, when they collaborate in an environment it can be viewed as an SoS. There is a large interest in modeling and simulation of phenomena present at different scales of SoS, which requires choosing appropriate model fidelity and resolution for the study. This thesis delves into the intricacies of modeling and simulation of mission scenarios within SoS, with a keen focus on engineered CS operating in physical environments. At the heart of the thesis lies the concept of model fidelity, examined through philosophical, scientific, and engineering perspectives. A maritime search and rescue scenario by airborne search assets serves as case study, using Agent Based Simulation (ABS) to explore the Measures of Effectiveness (MoE) of the mission. Through a hierarchical viewpoint it is studied how the model fidelity on Sub-System (SS) level affects its performance and how it aggregates up to performance on the CS level and at last the MoE on SoS level. In the scenarios, manned assets utilize traditional search patterns that is used in present search and rescue operations. Concurrently, autonomous assets employ a belief map search strategy where the map is based on probability density functions underpinned by a Bayesian modeling approach to new observations as the search progresses. The study shows that increasing the model fidelity at the SS level, particularly by refining the representation of the search sensor, can impact the MoE of the mission at the SoS level. However, this influence shows to be notably scenario-dependent, emphasizing the intricate relationship between model fidelity and system performance. The findings indicate that the relevance of model fidelity is tied to how measures of performance at lower levels aggregate to influence overarching mission outcomes. The thesis concludes by discussing validation and verification of the case study and ties encountered modeling difficulties to resolution and model fidelity aspects. Future work is suggested to be a deeper dive into mission engineering, where the mission itself is viewed as the system of interest, and how it could be combined with modeling and simulation in product development