NeuroGraph-PPO: A GNN-Based Proximal Policy Optimization Framework for Autonomous Power Routing in Mars Interplanetary Grids

This paper proposes a novel graph-reinforced learning approach forautonomous interplanetary power routing in Mars missions. We presentNeuroGraph-PPO, a hybrid framework combining Graph Neural Networks(GNN) and Proximal Policy Optimization (PPO) to optimize energy pathsin Martian microgrids. The model integrates outage forecasting, tariff prediction,load prioritization, and anomaly detection using machine learningand deep learning modules. A fully functional Python implementation isprovided and tested in simulated Mars-based grid environments.