Space habitation: Machine learning based evaluation and optimization of critical parameters affecting sustainable space habitats

Sustainable human habitation in space requires advanced predictive frameworks that can manage complex, interdependent systems. Current habitat models often rely on rule-based or simplified simulations, which inadequately capture the multifactorial interactions between environmental, physiological, and operational variables, limiting their effectiveness in long-duration missions. This study addresses this gap by applying machine learning (ML) techniques including Random Forest (RF), Support Vector Machine (SVM), Artificial Neural Networks (ANN), Gradient Boosting (GB), and Principal Component Analysis (PCA) to assess the impact of 27 critical parameters on space habitation sustainability. These parameters encompass environmental control, life support efficiency, psychological wellbeing, agricultural productivity, material degradation, and in situ resource utilization. Data from simulated space habitats and analog missions were processed to classify habitat performance into stability categories and identify high-impact variables influencing stability, such as atmospheric composition, temperature regulation, nutrient recycling, and social interaction dynamics. Results demonstrate that ensemble models outperform single learners, achieving 92.4 % predictive accuracy in correctly classifying habitat performance. Feature importance analysis reveals that oxygen-CO₂ balance, radiation shielding, social interaction quality, food production rate, and thermal stability are the top five determinants of habitation success. This research provides a scalable, data-driven framework to guide the design, control, and long-term sustainability of human life support systems in space habitation systems. Furthermore, the integration of these findings into future habitat designs could enhance resilience and adaptability, ensuring optimal living conditions for astronauts during extended missions. By leveraging these insights, future missions can be better equipped to tackle potential risks and uncertainties, ultimately fostering a more sustainable and effective approach to space habitation. This framework not only facilitates real-time monitoring and adjustments but also supports decision-making processes that prioritize astronaut health and mission success in the challenging environment of space.