Multi-rendezvous Spacecraft Trajectory Optimization with Beam P-ACO (1704.00702v1)

Abstract: The design of spacecraft trajectories for missions visiting multiple celestial bodies is here framed as a multi-objective bilevel optimization problem. A comparative study is performed to assess the performance of different Beam Search algorithms at tackling the combinatorial problem of finding the ideal sequence of bodies. Special focus is placed on the development of a new hybridization between Beam Search and the Population-based Ant Colony Optimization algorithm. An experimental evaluation shows all algorithms achieving exceptional performance on a hard benchmark problem. It is found that a properly tuned deterministic Beam Search always outperforms the remaining variants. Beam P-ACO, however, demonstrates lower parameter sensitivity, while offering superior worst-case performance. Being an anytime algorithm, it is then found to be the preferable choice for certain practical applications.

• The paper introduces a hybrid Beam P-ACO approach, integrating stochastic Beam Search with ant colony optimization to address multi-objective spacecraft trajectory design.
• It formulates trajectory planning as a bilevel optimization problem akin to the Traveling Salesman Problem, managing interplanetary dynamics effectively.
• Experimental results on the GTOC5 benchmark demonstrate enhanced adaptability and reduced tuning requirements, boosting overall mission efficiency.

Multi-rendezvous Spacecraft Trajectory Optimization with Beam P-ACO

The paper presents a novel approach to optimizing multi-rendezvous spacecraft trajectories using a hybrid of Beam Search and Population-based Ant Colony Optimization (P-ACO), termed as Beam P-ACO. This paper frames the spacecraft trajectory design as a multi-objective bilevel optimization problem, a task with significant complexity due to its combinatorial nature and the dynamic intricacies of interplanetary motion.

Key Components and Methodology

1. Problem Formulation: The trajectory design problem is a variant of the Traveling Salesman Problem (TSP) where the nodes are celestial bodies, and each arc represents a potential trajectory leg. The spacecraft must visit multiple bodies while minimizing the time and propellant mass required, thus maximizing the mission's scientific return.
2. Baseline Algorithm - Beam Search: Beam Search is employed as a baseline, where the search tree is pruned to allow only a limited set of promising paths, defined by a beam width, depth, and branching factor. This approach, however, is limited by its deterministic nature.
3. Stochastic Beam Search: Unlike traditional Beam Search, the stochastic variant introduces probabilistic decisions in the branching process, allowing exploration beyond deterministically pruned paths. This approach mitigates heuristic inaccuracies and enhances robustness.
4. Beam P-ACO Integration: The hybrid approach extends Stochastic Beam Search by integrating P-ACO principles. This involves successive generations of tree searches linked via pheromone trails that update anticipatory heuristics for node selection. Pheromones enable positive feedback, reinforcing paths associated with high-quality solutions, thus optimizing the search across generations.
5. Evaluation: The GTOC5 trajectory problem is utilized as the experimental benchmark. This problem engages an extensive database of asteroids, necessitating complex multi-body rendezvous maneuvers. The Beam P-ACO method is analyzed across various configurations of beam width and branching, focusing on achieving maximal mission scores, optimal propellant mass, and time efficiency.

Results and Implications

The experimental results reveal that while deterministic Beam Search can achieve high mission scores, it requires extensive parameter tuning. On the other hand, Beam P-ACO achieves comparable performance with less tuning by dynamically adapting to successful paths via pheromones. Additionally, Beam P-ACO demonstrates superior worst-case performance due to its inherent flexibility and adaptability through pheromone-guided search advancements.

Future Directions

The implications of these findings suggest that hybrid search algorithms like Beam P-ACO could be instrumental in optimizing complex trajectory problems where constraints are dynamically engaged. The research could further evolve by integrating additional decision-theoretic and adaptive strategies to enhance situational awareness and adaptability in aerospace mission planning.

Beam P-ACO holds promising potential for applications beyond space mission trajectories, including other domains where combinatorial and multi-objective optimization are required. Further advancements may involve exploring deeper integration with machine learning to adaptively learn and refocus search heuristics based on accumulated mission data, further enhancing trajectory planning efficiency.