Aplicación de la Red Generativa Adversarial condicional para la generación de soluciones arquitectónicas
Palabras clave:
modelos generativos; soluciones de diseño; zonificación espacialResumen
DOI: https://doi.org/10.46296/ig.v9i17.0344
Resumen
El uso de modelos generativo con inteligencia artificial dentro de la arquitectura está en un constante auge con un principal problema: generar variantes o diseños no siempre y normativamente construible. El trabajo de investigación implementa y evalúa el uso de la red generativa adversarial condicional como alternativa al uso de modelos generativo dentro de la arquitectura, con el objetivo de generar modelos a partir de una base de datos de 400 pares de planos arquitectónicos, los cuales contienen criterios de diseño en zonificación y distribución espacial de edificios residenciales. La implementación del modelo cGan reveló que las imágenes obtenidas poseen fidelidad con los planos de referencia, además de un aprendizaje progresivo y secuencial en su etapa de entrenamiento y generación de imágenes. El estudio también revela que el uso de estos modelos es ajustable a objetivos dependientes de cada proyecto, posicionándose como una herramienta que es capaz de generar diseños coherentes y optimizar el tiempo de modelación en tareas repetitivas y mecánicas en el desarrollo de planos de planta.
Palabras clave: modelos generativos; soluciones de diseño; zonificación espacial.
Abstract
The use of generative models with artificial intelligence in architecture is experiencing constant growth, with one main problem: generating design variants that are not always functionally or normatively feasible for construction. This research evaluates the use of the Conditional Generative Adversarial Network as an alternative to generative models in architecture, with the objective of generating models from a database of 400 pairs of architectural floor plans, which contain design criteria related to zoning and spatial distribution of residential buildings. The implementation of the cGAN model revealed that the generated images maintain fidelity to the reference floor plans, as well as demonstrating progressive and sequential learning during the training and image generation stages. The study also reveals that the use of these models can be adjusted to project-specific objectives, positioning itself as a tool capable of generating coherent designs and optimizing modeling time for repetitive and mechanical tasks in the development of floor plans.
Keywords: generative models; architectural layout generation; spatial planning.
Información del manuscrito:
Fecha de recepción: 09 de marzo de 2026.
Fecha de aceptación: 11 de mayo de 2026.
Fecha de publicación: 15 de junio de 2026.
Descargas
Citas
Jiang, H., Xiao, Y., Hurley, R., Liu, S. (2023). The Impact of Generative AI on Architectural Conceptual Design: Perfomance, Creative Self-Efficacy and Cognitive Load. https://doi.org/10.48550/arXiv.2601.10696
De Miguel Rodríguez, J., Villafañe M. E., Piškorec, L, Sancho Caparrini, F. (2020). Generation of geometric interpolations of building types with deep variational autoencoders. Design Science. https://doi.org/10.1017/dsj.2020.31
Zhong, C., Shi, Y., Cheung, L. H., Wang, L. (2024). AI-Enhaced performarmative building design optimization and exploration. Proceedings of the 29th International Conference of the Association for Computer-Aided Architectural Design Research in Asia (CAADRIA), 1, 59-68. https://doi.org/10.52842/conf.caadria.2024.1.059
Zhuang, J., Li, G., Xu, H., Xu, J., Tian, R. (2024). Text-to-City: Controllable 3D Urban Block Generation with Latent Diffusion Model. Proceedings of the 29th International Conference of the Association for Computer-Aided Architectural Design Research in Asia (CAADRIA), 2, 169-178. https://doi.org/10.52842/conf.caadria.2024.2.169
Zeng, P., Gao, W., Yin, J., Xu, P., Lu, S. (2024). Residential floor plans: Muti-conditional automatic generation using diffusion models. Automation in Construction, 162. https://doi.org/10.1016/j.autcon.2024.105374
Zhang, H., Zhang, R. (2025). Generating accesible multi-occupancy floor plans with fine-graind control using a diffusion model. Automation in construccion, 177. https://doi.org/10.1016/j.autcon.2025.106332
Karadag, I., Güzelci, O. Z., Alacam, S. (2023). EDU-AI: a twofold machine learning model to support classroom layout generation. Construction Innovation: Information Process Management, 23(4), 898-914. https://doi.org/10.1108/CI-02-2022-0034
Zhang, J., Fukuda, T., Yabuki, N., Li, Y. (2023). Synthesizing style-similar residential facade from semantic labeling according to the User-provided example. HUMAN-CENTRIC, Proceedings of the 28th International Conference of the Asociation for Computer-Aided Architectural Design Research in Asia (CAADRIA), 1, 139-148. https://doi.org/10.52842/conf.caadria.2023.1.139
Lin, H., Huang, L., Chen, Y., Zheng, L., Huang, M., Chen, Y. (2023). Research on the application of CGAN in the Design of historic Building Facades in Urban Renewal-Taking Fujian Putian Historic District as an example, China, Buildings, 13(6), 1478. https://doi.org/10.3390/buildings13061478
Gu, Y., Huang Y., Liao W., Lu, X. (2024). Intelligent design of shear wall layout based on diffusion models. Computer-Aided Civil and Infrastructure Engineering, 39, 3610-3625. https://doi.org/10.1111/mice.13236
Kösenciğ, K. Ö., Okuyucu, E. B., Balaban, Ö. (2024). Structural Plan Schema Generation Through Generative Adversarial Networks. Nexus Network Journal, 26, 409-427. 10.1007/s00004-024-00766-z
Leng, H., Gao, Y., Zhou, Y. (2024). ArchiDiffusion: A novel diffusion model connecting architectural layout generation from sketches to Wall Design. Journal of Building Engineering, 98. https://doi.org/10.1016/j.jobe.2024.111373
Li, C., Zhang T., Du, X., Zhang Y., Xie, H. (2025). Generative AI models for differente steps in architectural design: A literature review. Frontiers of Architecturl Research, 14(3), 759-783. https://doi.org/10.1016/j.foar.2024.10.001
Goodfellow, I. J., Pouget-Abadie, J., Mirza, M., Xu, B., Warde-Farley, D., Ozair, S., Courville, A., Bengio, Y. (2014). Generative Adversarial Nets. https://doi.org/10.48550/arXiv.1406.2661
Jabbar, A., Li, X., Omar, B. A. (2020). A Survey on Generative Adversarial Networks: variants, application, and training. ACM Computing Surveys (CSUR), 54(8), 1-49. https://doi.org/10.1145/3463475
Langr, J., Bok, V. (2019). GANs in Action: Deep learning with Generative Adversarial Networks. MANNING. https://dokumen.pub/gans-in-action-deep-learning-with-generative-adversarial-networks
Chaillou, S. (2019). ArchiGAN: Artificial Intelligece x Architecture. Architectural Intellugence: Selected Papers from the 1st International Conference on Computational Design and Robotic Fabrication (CDRF), p. 117-127, 2019. https://papers.cumincad.org/data/works/att/artificial_intellicence2019_117.pdf
Shi, Y.; Hong, T. C. (2026). Benchmarking pix2pix on floor Plans: Shape Grammar as a Deterministic Standard for Models Limits. In Conference Paper: CAADRIA. https://www.researchgate.net/publication/403581693_Benchmarking_pix2pix_on_Floor_Plans_Shape_Grammar_as_a_Deterministic_Standard_for_Model_Limits
Guo, Y., Fang, T., Cui, Z., Stouffs, R. (2025). A dual-aspect evaluation framework for architectural-like plan generation via pix2pix series algorithms. Frontiers of Architectural Research, 14 (6), 1516-1535. https://doi.org/10.1016/j.foar.2025.04.006
Isola, P., Yan-Zhu, J., Zhou, T., Efros, A. A. (2018). Image-to-Image Translation with Contional Adversarial Networks. https://doi.org/10.48550/arXiv.1611.07004
Центр методологии нормарования и стандартизации в строительстве (ФГУП ЦНС), (2005). СП 31-107-2004: Архитектурно-Планировочные решения многоквартирных жилых зданий. Москва. https://ohranatruda.ru/upload/iblock/9de/4294813059.pdf
Центральный научно-исследовательский и проектно-экспериментальный институт промышленных зданий и сооружений. (2022). СП 118.13330.2022: Общественные здания и сооружения. https://rkc56.ru/attach/orenburg/docs/Gosstandart_RF/sp_118_13330_2022_.pdf
Министерство строительства и жилишно-коммунального хозяйства российской федерации. (2022). СП 54-13330.2022: Здания жилые многоквартирные. Москва. https://rkcson-kbr.ru/images/docs/dostup_sreda/sp_54133302022.pdf
Abadi, M., Barham, P., Chen, J., Chen, Z., Davis, A., Dean, J., Devin, M., Ghemawat, S., Irving, G., Isard, M., Kudlur, M., Levenberg, J., Monga, R., Moore, S., Murray D. G., Steiner, B., Tucker, P., Vasudevan, V., Warden, P., Wicke, M., Yu, Y., Zheng, X. (2016). TensorFlow: A system for large-scale machine laerning. https://doi.org/10.48550/arXiv.1605.0869
Cira, C., Manso-Callejo, M., Alcarria, R., Pareja, T. F., Sanchez, B. B., Serradilla, F. (2021). Generative Learning for Postprocessing Semantic Segmentation Predictions: A Lightweight Conditional Generative Adversarial Network Based on Pix2Pix to Improve the Extraction of Road Surface Areas, buildings, 10(1). https://doi.org/10.3390/land10010079
Betker, J., Goh, G., Jing, L., Brooks, T., Wang, J., Li, L., Ouyang, L., Zhuang, J., Lee, J., Guo, Y., Manassra, W., Dhariwal, P., Chu, C., Jiao, Y. (2023). Improving Image Generation with Better Captions. https://cdn.openai.com/papers/dall-e-3.pdf
Imagen 3 Team Google (2024). Imagen 3. https://doi.org/10.48550/arXiv.2408.07009
Horvath, A.S., Pouliou, P. (2024). AI for conceptual architecture: Reflections on designing with text-to-text, text-to-imagen, and image-to-image generators. Frontiers of Archiectural Research, 13(3), 593-612. https://doi.org/10.1016/j.foar.2024.02.006
Chen, J., Wang, D., Shao, Z., Zhang, X., Ruan, M., Li, H., Li, J. (2023). Usign Artificial intelligence to generate Master-Quality Architectural Designs from text Descriptions. buildings, 13(9), 2285. https://doi.org/10.3390/buildings13092285
Rodriguez, R. C., Duarte, R. B. (2022). Generating floor plans with deep learning: A croos-validation assessment over different dataset sizes. International Journal of Archutectural Computing, 20(3), 630-644. https://doi.org/10.1177/14780771221120842
Nadari, M., Karami, N., Emami, A., Shirani, S., Samavi, S. (2022). Dynamic-Pix2Pix: Noise Injected cGan for Modeling Input and Target Domain Joint Distributions with Limited Trainig Data. https://doi.org/10.48550/arXiv.2211.08570
Publicado
Cómo citar
Número
Sección
Licencia
Derechos de autor 2026 Revista Científica INGENIAR: Ingeniería, Tecnología e Investigación. ISSN: 2697-3693.
Esta obra está bajo una licencia internacional Creative Commons Atribución-NoComercial-CompartirIgual 4.0.
