近年來，在建築設計方面導入人工智慧的案例越來越多，由參數模型、自動化衍生設計、建築性能模擬和目標優化組成，已成為輔助建築師設計的有效工具，本研究探索設計初期階段，利用數位工具輔助自動化量體草案衍生的最佳化方案，以及應用在建築設計實務流程的可能性。
現有的設計方法流程為透過基地分析（如：動線分析、植栽分析、環境分析、建蔽率、容積率及高度限制）後，由人工生成多種配置方案後，進行對比並選擇其中一種方案後進行修改，並在最後進行環境模擬分析。傳統流程的缺點在於線性作業，無法同時配置多種方案，較容易陷入單一設計思維中，無法模擬一同列入配置考量。本研究聚焦在建築設計初期階段的設計方法流程，以建築初期草案自動化衍生為目標，整合衍生式工具進行最佳化之建築設計方案自動生成。
本研究結合Rhino Grasshopper及Revit應用，將基地分析的限制條件參數化，連接量體生成模組，並將環境模擬提前加入參數應用，透過優化生成多種方案配置後挑選修改，在利用Revit inside Rhino導入Revit中進建築資訊模型整合，以利於設計初期草案之進行。
本研究以成功大學旺宏館設計案進行測試，透過參數調整來設定基地限制，加入植栽保留區域、預期預留空地區域及建蔽率、總樓地板面積、樓層高度等限制，透過grasshopper插件Evomass來生成建築量體，以參數設定來限制量體生成的範圍、量體個數及單元量體的最大最小值，此外，並結合建築性能模擬，設定優化目標，透過懲罰或獎賞進行運算優化，此優化算法利用Evomass中的SSIEA進行優化運算，生成符合基地限制、設計者需求及性能模擬目標的量體草案配置。本研究透過案例驗證，標的多種量體方案，再由設計師進行挑選後修改。修改後的量體可透過Revit inside Rhino插件，將原有在Rhino生成之量體於Revit中創建可供建築資訊模型（BIM）使用之模型，達到與BIM平台的結合，以利後續在建築設計上BIM協同作業的進行。
本研究實證的結果，透過建築師訪談、回饋與分析，顯示加入自動化草案生成後的建築初期設計方法流程，可優化生成較佳的方案或是提供新的設計想法，因提前加入環境模擬優化，可在建築設計初期階段快速提供可行性草案分析，使建築師有更多的時間能投入細部設計及修改方案，可作為未來建築師事務所導入人工智慧共同協作之參考。

Generative design, parametric modeling, building information modeling (BIM), and building performance optimization have been used in architectural design to solve design complexity. This study uses generative design to create a novel automatic architectural design process in the early stages of architecture. It has provided a potential for solving complex design problems and improving work efficiency.

中文文獻
[1] 王力凱. (2020). 基於性能的建築量體設計 (Doctoral thesis, 南京大學).
[2] 簡聖芬. (1999). 衍生式設計輔助系統:現況與未來發展.
[3] 胡一波, & 王宇平. (2009). 解約束優化問題的一種新的罰函數模型. 計算機科學, 36(7), 240-243.
[4] 李俊儀. (2001). 衍生式與演化式建築輔助設計系統.
[5] 胡一波, & 王宇平. (2009). 解約束優化問題的一種新的懲罰函數模型. 計算機科 學, 36(7), 240-243.
英文文獻
[1] Asimow, M. (1962). Introduction to Design. Englewood Cliffs, NJ: Prentice-Hall.
[2] Alexander, C. (1964). Notes on the Synthesis of Form. Cambridge, MA: Harvard University Press.
[3] Coyne, R. & U. Flemming. (1990). "Planning in design synthesis: abstraction-based LOOS." Artificial Intelligence in Engineering. Vol. 1, J. Gero, ed., New York: Springer, 91-111.
[4] Chien, S.-F. (1998). "Supporting information navigation in generative design systems." Ph.D. Dissertation, School of Architecture, Carnegie Mellon University, Pittsburgh, PA.
[5] Costa, L. M. D., Alvarez, C. E. D., & Martino, J. A. D. (2021). Proposta de método de projeto baseado no desempenho para edifícios energeticamente eficientes. Ambiente Construído, 21, 409-433.
[6] Duering, S., Chronic, A., & Koenig, R. (2020, May). Optimizing Urban Systems: Integrated optimization of spatial configurations. In Proceedings of the 11th Annual Symposium on Simulation for Architecture and Urban Design (pp. 1-7).
[7] Duering, S., Chronic, A., & Koenig, R. (2020, May). Optimizing Urban Systems: Integrated optimization of spatial configurations. In Proceedings of the 11th Annual
83
Symposium on Simulation for Architecture and Urban Design (pp. 1-7).
[8] Eastman, C.M. (Ed.) (1975). Spatial Synthesis in Computer- Aided Building Design. London, England: Applied Science Publishers.
[9] Flemming, U. & R. Woodbury. (1995). "Software environment to support early phases in building design (SEED): overview." Journal of Architectural Engineering 1, 4, 147- 152.
[10] Gero, J. (1977). Computer Applications in Architecture. London, England: Applied Science Publishers.
[11] Gannon, M. (2017). Human-centered Interfaces for autonomous fabrication machines.
[12] Heisserman, J. (1991). "Generative geometric design and boundary solid grammars." Ph.D. Dissertation, Department of Architecture, Carnegie Mellon University, Pittsburgh, PA.
[13] Mitchell, W.J. (1977). Computer-Aided Architectural Design. New York, NY: Van Nostrand Reinhold.
[14] Newell, A. & H.A. Simon. (1972). Human Problem Solving. Englewood Cliffs, NJ: Prentice-Hall.
[15] Narangerel, A., Lee, J. H., & Stouffs, R. (2017). Thermal and Daylighting Optimization of Complex 3D Faceted Façade for Office Building.
[16] Nauata, N., Chang, K. H., Cheng, C. Y., Mori, G., & Furukawa, Y. (2020, August). House-gan: Relational generative adversarial networks for graph-constrained house layout generation. In European Conference on Computer Vision (pp. 162-177). Springer, Cham.
[17] Veloso, P., & Krishnamurti, R. (2020). An Academy of Spatial Agents-Generating spatial configurations with deep reinforcement learning.
[18] Waibel, C., Bystricky, L., Kubilay, A., Evins, R., & Carmeliet, J. (2017). Validation of Grasshopper-based Fast Fluid Dynamics for Air Flow around Buildings in Early Design Stage. In Building Simulation (pp. 7-9).
[19] Wang, L., Janssen, P., & Ji, G. (2019). Reshaping Design Search Spaces for Efficient Computational Design Optimization in Architecture. In ICCC (pp. 100-107).
[20] Wortmann, T. (2017). Model-based optimization for architectural design: Optimizing daylight and glare in grasshopper. Technology| Architecture+ Design, 1(2), 176-185.
[21] Wortmann, T. (2019). Architectural Design Optimization—Results from a User Survey. KnE Social Sciences, 473-483.
[22] Wortmann, T. (2019). Genetic evolution vs. function approximation: Benchmarking algorithms for architectural design optimization. Journal of Computational Design and Engineering, 6(3), 414-428.
[23] Wortmann, T., & Schroepfer, T. (2019, April). From optimization to performance- informed design. In Proceedings of the Symposium on Simulation for Architecture and Urban Design(pp. 1-8).
[24] Wang, L., Chen, K. W., Janssen, P., & Ji, G. (2020). Enabling Optimisation-based Exploration for Building Massing Design-A Coding-free Evolutionary Building Massing Design Toolkit in Rhino-Grasshopper.
[25] Wang, L., Chen, K. W., Janssen, P., & Ji, G. (2020). Algorithmic generation of architectural Massing Models for building design optimisation-Parametric Modelling Using Subtractive and Additive Form Generation Principles.
[26] Wang, L., Janssen, P., & Ji, G. (2020). SSIEA: a hybrid evolutionary algorithm for supporting conceptual architectural design. AI EDAM, 34(4), 458-476.
[27] Wortmann, T., & Fischer, T. (2020). Does Architectural Design Optimization Require Multiple Objectives?-A critical analysis.
[28] Yu, D. (2020). Reprogramming Urban Block by Machine Creativity-How to use neural networks as generative tools to design space.
[29] Zheng, H., An, K., Wei, J., & Ren, Y. (2020). Apartment Floor Plans Generation via Generative Adversarial Networks.
[30] Zheng, H., & Ren, Y. (2020). Machine Learning Neural Networks Construction and Analysis in Vectorized Design Drawings.
[31] Zhang, H., Wang, L., & Ji, G. (2021). The Synergy of building massing and facade-An Evo-Devo approach for performance-based design optimization combining facade design with building massing.
網路文獻
[32] Ladybug-Grasshopper / Avaible online : https://www.ladybug.tools/
[33] Ladybug, Food4Rhino / Avaible online : https://www.food4rhino.com/en/app/ladybug-tools
[34] ClimateStudio / Avaible online : https://www.solemma.com/climatestudio
[35] GH_Wind / Avaible online : https://github.com/christophwaibel/GH_Wind
[36] Wikipedia, 標準氣象年(TMY3) / Avaible online : https://zh.wikipedia.org/wiki/%E6%A8%99%E6%BA%96%E6%B0%A3%E8%B1%
A1%E5%B9%B4
[37] 蘇建儒 (2019). / Avaible online : http://www.taisuco.com/monthly/CPN.aspx?ms=1446&p=13387451&s=13387473
[38] 何謂BIM / Avaible online : https://sites.google.com/site/bimzixunpingtai/home/he-weibim
[39] Jianyouli (2014). 何謂衍生式設計 Avaible online :
[40] 什麼是參數化設計? / Avaible online : https://jianyouli.wordpress.com/2014/02/01/%E4%BD%95%E8%AC%82%E8%A1% 8D%E7%94%9F%E8%A8%AD%E8%A8%88-generative-design%EF%BC%9F/
[41] Kean Walmsley (2019). / Avaible online : https://www.keanw.com/2019/11/project-rediscover-version-2-of-the- mars-graph-is-now-available-for-download.html
[42] Wikipedia (2021). Rhinoceros3D / Avaible online : https://zh.wikipedia.org/wiki/Rhinoceros_3D
[43] 陳珍誠 (2017). 參數化設計:建築設計中的演算思維 / Avaible online :
http://www.twarchitect.org.tw/special/%E5%8F%83%E6%95%B8%E5%8C%96%E8 %A8%AD%E8%A8%88%EF%BC%9A%E5%BB%BA%E7%AF%89%E8%A8%A D%E8%A8%88%E4%B8%AD%E7%9A%84%E6%BC%94%E7%AE%97%E6%80 %9D%E7%B6%AD/?doing_wp_cron=1638792667.1930348873138427734375
[44] 什麼是參數化設計 / Avaible online : https://3dp-maker.blogspot.com/2019/07/parametric-modelling-and- Direct-Modeling-SOLIDWORKS-AutodeskInventor-CreoParametric-CA TIA-NX- nshape.html
[45] Wallacei-Grasshopper / Avaible online : http://www/wallacei.com/
[46] Green BIM 微氣候資訊平台 / Avaible online : http://www.weatherservice.org.tw/
[47] Rhino.Inside.Revit / Avaible online : https://www.rhino3d.com/tw/features/rhino-inside-revit/
[48] 適應度與選擇,2019 / Avaible online : https://blog.csdn.net/weixin_30239361/article/details/101540896
[49] 陽傘可視化圖表生成教學 / Avaible online : https://github.com/ParasolJS/parasol-es/wiki/Tutorial-1:-Getting- started