DESIGN OF CISLUNAR ON-ORBIT SERVICING NETWORKS

This dissertation develops an integrated framework for designing satellite servicing networks in the Earth–Moon region, with the goal of making refueling, repair, and repositioning of spacecraft in cislunar space efficient and sustainable. The work treats servicing as a coupled problem in three parts, orbit selection, transfer design, and schedule design, that must be solved together rather than in isolation. First, it introduces a procedure to curate a manageable set of candidate orbits for hubs, servicer vehicles, and customer spacecraft, based on access, station-keeping, deployment cost, coverage, and how well different orbits work together as a network rather than as individual “best” orbits.Second, the research develops a phasing-informed transfer design pipeline that uses the natural multi-body dynamics of the Earth–Moon system to seed long-range rendezvous options, primarily through invariant manifold arcs, then refines selected candidates with differential corrections. This process produces heatmaps and cost surfaces that capture both ΔV and time-of-flight as functions of phasing, and it includes extensions for nearly stable orbits, clustering of trajectory “motion primitives”, and ephemeris-model validation so that designers can quickly assemble credible transfer options without exhaustively converging every trajectory.Third, the dissertation formulates servicing as an orbit traveling salesperson problem on a time-dependent graph, and develops two complementary scheduling approaches built on the dynamics-consistent transfer library. An auction-based algorithm provides fast, scalable assignments suitable for rapid updates and potential onboard use, while a tailored genetic algorithm explores ordering, phasing choices, optional tasks, and even orbit selection, to find lower-ΔV campaigns at higher computational cost. A pre-seeded auction variant narrows the performance gap with the genetic algorithm while cutting runtime by roughly an order of magnitude, making it attractive for real-time or on-orbit re-planning.Integrated case studies across planar, spatial, and resonant orbit families, including an Earth–Moon–Sun ephemeris demonstration, show how this unified process can select orbits, construct transfers, and build executable schedules that trade fuel, response time, and fleet size for different mission priorities. Overall, the research provides both methods and design guidance for government and commercial stakeholders who need to plan, compare, and field cislunar on-orbit servicing capabilities at the campaign level.