Climate-Adaptive Generative Design of Residential Buildings Using Reinforcement Learning and Parametric Modeling
Keywords:
reinforcement learning; parametric design; climate adaptation; residential buildings; multi-objective optimization; building performance simulationAbstract
Residential building design under climate uncertainty requires simultaneous decisions about envelope geometry, glazing, shading, thermal mass, and natural ventilation, yet these decisions are usually made before operational control strategies are known. This paper develops a methodological framework for climate-adaptive generative design in which reinforcement learning learns a policy over parametric residential design variables rather than only over HVAC setpoints. The design problem is formulated as a constrained multi-objective Markov decision process in which states encode partial design configurations and simulation-derived performance, actions modify geometric and envelope parameters, and rewards combine normalized energy, thermal discomfort, life-cycle carbon, and cost terms with feasibility penalties. A parametric modeling layer maps design actions to simulation-ready geometry and building systems, while a simulation layer evaluates performance across climate scenarios and occupancy assumptions. Proximal policy optimization is used to stabilize learning under high-dimensional, non-stationary design rewards, and a robust evaluation rule integrates expected and worst-case climate performance. The framework also specifies Pareto filtering, constraint handling, uncertainty treatment, and sensitivity analysis so that generated designs can be assessed as trade-off candidates rather than as a single optimum. The main contribution is the formal coupling of generative design with a learned sequential decision policy, enabling the production of diverse, climate-conditioned residential design candidates and providing a technically defensible basis for early-stage design exploration. Practical significance lies in linking envelope-level climate adaptation with operational flexibility and in making constraint-aware design search more scalable than repeated independent optimization runs.
References
1. Crawley, D. B., Lawrie, L. K., Winkelmann, F. C., Buhl, W. F., Huang, Y. J., Pedersen, C. O., Strand, R. K., Liesen, R. J., Fisher, D. E., Witte, M. J., & Glazer, J. (2001). EnergyPlus: creating a new-generation building energy simulation program. Energy and Buildings, 33(4), 319–331.
2. Loonen, R. C. G. M., Trčka, M., Cóstola, D., & Hensen, J. L. M. (2013). Climate adaptive building shells: State-of-the-art and future challenges. Renewable and Sustainable Energy Reviews, 25, 483–493.
3. Attia, S., Hamdy, M., O'Brien, W., & Carlucci, S. (2013). Assessing gaps and needs for integrating building performance optimization tools in net zero energy buildings design. Energy and Buildings, 60, 110–124.
4. Wei, T., Wang, Y., & Zhu, Q. (2017). Deep reinforcement learning for building HVAC control. In Proceedings of the 54th Annual Design Automation Conference (pp. 1–6). ACM.
5. Zhang, Z., Chong, A., Pan, Y., Zhang, C., & Lam, K. P. (2019). Whole building energy model for HVAC optimal control: A practical framework based on deep reinforcement learning. Energy and Buildings, 199, 472–490.
6. Azuatalam, D., Lee, W. L., de Nijs, F., & Liebman, A. (2020). Reinforcement learning for whole-building HVAC control and demand response. Energy and AI, 2, 100020.
7. Krish, S. (2011). A practical generative design method. Computer-Aided Design, 43(1), 88–100.
8. Tabasi, S. F., Luo, D., Rafizadeh, H. R., Le Nguyen, K., & Banihashemi, S. (2025). Generative optimization of building blocks for density, solar and structural performance. Journal of Building Engineering, 111, 113307.
9. Evins, R. (2013). A review of computational optimisation methods applied to sustainable building design. Renewable and Sustainable Energy Reviews, 22, 230–245.
10. Machairas, V., Tsangrassoulis, A., & Axarli, K. (2014). Algorithms for optimization of building design: A review. Renewable and Sustainable Energy Reviews, 31, 101–112.
11. Tian, W., Heo, Y., de Wilde, P., Li, Z., Yan, D., Park, C. S., & Augenbroe, G. (2018). A review of uncertainty analysis in building energy assessment. Renewable and Sustainable Energy Reviews, 93, 285–301.
12. Sutton, R. S., & Barto, A. G. (2018). Reinforcement learning: An introduction (2nd ed.). MIT Press.
13. Schulman, J., Wolski, F., Dhariwal, P., Radford, A., & Klimov, O. (2017). Proximal policy optimization algorithms. arXiv. https://arxiv.org/abs/1707.06347
14. Dulac-Arnold, G., Mankowitz, D., & Hester, T. (2019). Challenges of real-world reinforcement learning. arXiv. https://arxiv.org/abs/1904.12901
15. Negendahl, K. (2015). Building performance simulation in the early design stage: An introduction to integrated dynamic models. Automation in Construction, 54, 39–53.
16. Ostergård, T., Jensen, R. L., & Maagaard, S. E. (2016). Building simulations supporting decision making in early design – A review. Renewable and Sustainable Energy Reviews, 61, 187–201.
17. Reinhart, C. F., & Cerezo, C. (2016). Urban building energy modeling – A review of a nascent field. Building and Environment, 97, 196–202.
18. Nguyen, A. T., Reiter, S., & Rigo, P. (2014). A review on simulation-based optimization methods applied to building performance analysis. Applied Energy, 113, 1043–1058.
Downloads
Published
Issue
Section
License
Copyright (c) 2026 International Journal of Artificial Intelligence Engineering and Systems
This work is licensed under a Creative Commons Attribution 4.0 International License.
This article is published under the Creative Commons Attribution 4.0 International License (CC BY 4.0), which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited.
