Ph.D. Proposal: Tristan Sarton du Jonchay

Fri Oct 22 2021 04:00 PM
“Space Logistics Optimization Framework for the Design and Operations of Sustainable On-Orbit Servicing Infrastructures”

You're invited to attend


Ph.D. Proposal


Tristan Sarton du Jonchay

(Advisor: Prof. Koki Ho)


“Space Logistics Optimization Framework for the Design and Operations of Sustainable On-Orbit Servicing Infrastructures”


Friday, October 22
4:00 p.m.


On-Orbit Servicing (OOS) is a nascent space-based industry aimed at making the operations and management of Earth-orbiting satellites sustainable. Two space systems central to OOS infrastructures are servicers and orbital depots. Servicers are robotic spacecraft designed to provide a set of services to client satellites, such as refueling and/or repair. Orbital depots are in-space warehouses that store commodities such as spares and/or propellant to support the long-term operations of the servicers and client satellites. This work develops an OOS logistics planning framework that models and optimizes the operations of complex on-orbit servicing infrastructures involving several servicers and depots that provide multiple types of services to a fleet of Earth-orbiting satellites. The framework contributes in three different ways to the problem of OOS planning: (1) OOS operators can optimally manage their infrastructures on a daily to monthly basis; (2) OOS decision makers can design prospective infrastructures by assessing their long-term value for a given state of the servicing market; and (3) a large tradespace related to the flights of the servicers enables realistic OOS operations and a large set of design options available to decision makers. The first contribution is achieved by extending state-of-the-art space logistics techniques – traditionally used for the optimal design of long-term, large-scale human space exploration missions – to the unique problem of OOS logistics and service management. The logistics of OOS (e.g., allocation of commodities to servicers) is modeled as a Time-Expanded Dynamic Generalized Multi-Commodity Network Flow problem and optimally solved as a Mixed-Integer Linear Program. The provision of services to client satellites is uniquely modeled as a binary problem and added to the logistics formulation. The second contribution is achieved by applying the Rolling Horizon procedure to the OOS problem. This technique weaves service demand uncertainty into the OOS optimization framework. This allows decision makers to design servicing infrastructures that are competitive over the long term by capitalizing on a broad range of deterministic and random service needs. In addition, this procedure improves the computational efficiency of OOS problems defined over long time horizons by breaking them down into smaller subproblems which are solved consecutively and whose combined optimal solutions yield a satisfying solution to the original problems. The third contribution is achieved by: (a) modeling the relative dynamics of the nodes of the OOS logistics network; (b) modeling the optimal trajectories of high-thrust servicers, low-thrust servicers, and multimodal servicers (i.e., equipped with both low-thrust and high-thrust propulsion systems); and (c) using Gaussian Processes to build surrogates of the computationally expensive trajectory optimization models.


  • Dr Koki Ho – School of Aerospace Engineering (advisor)
  • Dr Glenn Lightsey – School of Aerospace Engineering
  • Dr Sandra Magnus – School of Mechanical Engineering
  • Dr Mark Whorton – Chief Technology Officer, Georgia Tech Research Institute