進年來，隨著數字化技術的普及和多學科思維模式的整合，建築自動生成設計作為一種建築學前沿的設計方法逐步進入人們的視野。而對於小住宅基地來說，設計前期通常需要不斷通過對法規、建築量體、戶型及配套設施進行反復的調整和指標計算來滿足居住環境的功能要求。因為小住宅基地的特性，設計前期工作量大、設計所呈現的牽一髮而動全身的特性也嚴重影響了設計推動效率、實施者與地主的協商時間漫長所致建築師大多不願積極投入。
因此，對於實施者能有初步建築設計方案，供雙方協商的討論圖說，即成為實施者殷切盼望的工具。故而，自動生成建築平面設計多種方案，即成為數位化建築設計之研究動機。研究的主要目的是利用參數化設計對建築基地進行各層平面初步設計，運用參數調控快速的產生多個可行性方案供實施者協商之參考。研究過程將設計過程拆分成若干步驟進行程序化。將基礎法規轉換為操作過程之限制條件與電腦運算思維結合，創造出更多的設計可行性。
本文利用空間之間的功能拓撲關係實現了空間與參數化設計運算方式之間的連動和生成，通過模型的生成邏輯使用Grasshopper 來搭建起一系列模塊化的建模指令，實現3D 模型的實時修改和可視化展現的生成方案。從而使本文實現了在小基地內自動生成平面；在建築平面的可視性和實時聯動修改法規限制、基地大小、平面空間關係連結上更為直觀和方便為非專業者使用操作；對生成平面的優化篩選提供了不同的條件選擇，更智慧、快速地為不同的小住宅基地提供多元的選擇方案。本文清晰的總結設計步驟和工作流程，以區分人腦與電腦分工的格局；嘗試用程式代替機械製圖；也對人工智能介入建築空間佈局設計做出了進一步的展望。
For architectural graphic design, the regulatory requirements of the building base and the early stage of graphic design are relatively time-consuming and laborious stages. Professional designers are required to design and modify multiple schemes based on experience to determine a suitable graphic scheme. The process lacks innovation and the computational workload is huge. Due to the numerous regulatory requirements for urban renewal of dangerous old buildings and complex base requirements. So choose it as the main object of research.
The main purpose of this article is to use parametric design to conduct a preliminary plane design for the old and dangerous building base, to intelligently and quickly generate multiple feasible plans and use the building physical environment analysis to screen out the best plan Split the design process into several steps for proceduralization.Turn artificial intelligence into a tool for architects, save time, improve design efficiency, and reduce error rates.Combine the abstract thinking of the human brain with the computational thinking of the computer to create more design possibilities. In recent years, architectural design has made rapid progress in artificial intelligence and commercialization.
This article uses the functional topological relationship between spaces to achieve linkage and generation between space and parametric design calculation methods as an entry point. Through the model generation logic, Grasshopper is used to build a series of modular modeling instructions to realize the real-time real-time modification and visualization of the 3D model. After generating the plan, multiple indicators are evaluated and selected based on the evolutionary algorithm. Based on the above strategies, this research explored the possibility of automatically generating a plane in a small base and obtained the following results: The visibility of the building plane and real-time linkage modification of legal restrictions, the size of the base, and the connection of plane spatial relations are more intuitive and It is convenient for non-professionals to use and operate. It provides different conditions for the optimization and selection of generated planes, and provides intelligent and fast selection schemes for dangerous and old buildings at different bases; Clear summary of design steps and Work flow to distinguish the division of labor between the human brain and the computer; Try to replace mechanical drawing with programs, and also look forward to artificial intelligence intervention in architectural space layout design.
