Multi-Objective Graph Search for Automated Planetary Moon Tour Design

This paper proposes and tests a novel strategy to automate the flyby sequence design of planetary moon tours. The v∞ globe is used to generate a database of families of ballistic transfers and V-Infinity Leveraging Maneuvers (VILTs) in the Keplerian 0-Sphere of Influence model, thus discretizing the design space. The database is then used to translate the trajectory optimization problem into a graph path-planning problem by leveraging a new tool introduced here, the Hopping Graph. The graph is built by considering all the possible flybys linking two transfers as nodes, while the edges are the transfers between two flybys. Edge weights are defined as a function of the characteristics of the transfers they represent. The main advantage of using a graph formulation is that, by applying exact algorithms such as Dijkstra’s, it is guaranteed that the optimal solution is obtained for the considered design space, with respect to the selected edge-weighting function. The Hopping Graph framework is here applied to the Endgame moon tour problem, using the Dijkstra algorithm on the graph with a weighted sum of transfer ΔV and time of flight as the edge-weighting function. Additionally, the multiobjective EMOA∗ algorithm is also used to obtain the Pareto front of the shortest paths. A relevant case study in the Saturnian system is presented, showing improvements with respect to the existing literature on Endgame moon tour design, both in terms of results and computational times.