Stochastic Mission-Sampling for Comparing On-Orbit Servicer Propulsion Architectures Under Uncertainty

Selecting propulsion architectures for on-orbit servicing requires balancing responsiveness, efficiency, and adaptability under an uncertain servicing tempo in both low Earth orbit (LEO) and geostationary orbit (GEO). To address this uncertainty, a mission-sampling framework is introduced to evaluate multiple propulsion architectures for on-orbit servicing. The approach employs a mission-sequence Monte Carlo sampler in conjunction with General Mission Analysis Tool propagation to generate stochastic servicing campaigns, enabling quantitative assessment of mission performance, within-policy flexibility, cross-policy steerability, and robustness. Servicing requests are sampled from a fixed categorical demand model, representing a memoryless Markov chain, with the option to extend to state-dependent transition matrices for clustered or time-varying demand. Four propulsion architectures—chemical, electric, hybrid (split-tank), and single-propellant multimode—are evaluated using interpretable metrics, including geometric outcome dispersion, mode compliance, and time-integrated [Formula: see text] capability envelopes. Results demonstrate that the multimode architecture provides substantially larger capability envelopes ([Formula: see text] in GEO and [Formula: see text] in LEO) and greater aggregate flexibility ([Formula: see text] in GEO and [Formula: see text] in LEO), with full mode compliance under mixed operational policies, compared to hybrid architectures. In contrast, hybrid configurations are limited by early propellant depletion and experience significant performance losses when fixed allocations are committed, resulting in reduced operational capability.