Designing Space-Based Architectures to Discover and Track Hazardous Asteroids Using MO-MCTS

Abstract The discovery and tracking of near-Earth objects (NEOs), such as asteroids and comets, form the foundation of any effective planetary defense mitigation strategy—that is, the process of reducing or eliminating the threat these objects pose to humanity. Establishing sufficient warning time prior to a potential Earth impact is a critical outcome of this capability, as it enables informed decision-making and the selection of viable mitigation options. This study investigates how to optimally pre-position a fleet of NEO discovery and tracking (D&T) spacecraft within the Sun-Earth system. Candidate architectures—one or more pre-positioned spacecraft—are designed using Multi-Objective Monte Carlo Tree Search (MO-MCTS). Architecture performance is evaluated using a target deck of one thousand synthetic asteroids. The resulting designs exhibit trade-offs among warning time, cost, and remote-sensing coverage. For single-spacecraft architectures, the highest-performing designs employ distant retrograde orbits (DROs) as well as L2 and L3 vertical orbits. Among multi-spacecraft architectures, the most frequently selected orbit families include L3 halo orbits, L4 and L5 short- and long-period orbits, and those orbits identified as high-performing in the single-spacecraft case.