早期設計階段是一種強調設計生產（design production）的過程（Goldschmidt 1995）。 設計生產為結合二人以上的參與者進行互動性的討論，經由集體產生大量且多樣性的 設計想法，予以組織（篩選、組合或調整）而發展概念設計用以解決設計問題。連結 想法為設計生產過程的主要機制，參與者不僅連結自身內在的記憶，同時也連結其他 參與者的知識，其集體建構圖像般（graph-like）的連結知識可視為共同的設計空間。 由於國際化與節能化的趨勢，此設計生產過程將在分散性設計環境中進行。 本計劃以從事建築設計的團隊（如建築師事務所、建築設計工作室等）為例，探討參 與者在早期設計階段的設計生產過程中，如何連結想法並組織此連結知識用以解決設 計問題，且藉由對運算理論的相關研究，包括設計空間探索 (design space exploration)、 輔助分散式知識連結、與連結知識建構與呈現，探索其可能的運算機制與設計理論模 型。 研究方法乃透過運算方式進行下列的步驟，包括1)實驗觀察與記錄、資料蒐集、分析 與整理；2) 核心理論分析系統建構；3) 設計實驗進行系統檢驗。藉由此研究方法，並 結合先前建構之輔助分散式想法連結系統(DIM)(Lai 和 Chang, 2004)，本計畫企圖建構 一個設計決策支援系統雛形，用以輔助設計團隊在分散式的設計生產過程中，能有效 的解決設計問題，用以提昇國內建築設計在產業與學界的競爭力。 The early design stage focuses on the process of design production (Goldschmidt 1995). For solving the design problem, two more participants often have the interactive meeting to generate diverse ideas and then organize (combining or adapting) the ideas to develop conceptual designs. Linking ideas is the key mechanism in the process of design productive. A designer links ideas through his/her individual memory as well as other participants’ knowledge. Consequently a graph-like knowledge structure, which can be regarded as the group design space, is constructed by the participants. Due to the tendency of internationalization and energy saving, the process of design productive will be executed in the distributed design environment. The main observation targets are architectural design teams due to their majority representing the current industry and education besides their design-oriented capability needed by the nature of this research. Through understanding how participants link ideas and use the graph-like knowledge structure in the process of design productive, this research explores the feasibility of computational model as well as a formal design model of the domain of interest. In addition, three computational theories are investigated to study related computational mechanisms including design space exploration, representation and organization of linking knowledge, and supporting distributed linking. We apply the computational methodology to approach this research. Three steps are 1) design experiment: data recording and analyzing the graph-like knowledge structures; 2) literature reviews and theory analysis: according to the result of analysis; 3) system implementation and testing: implementing a prototype for supporting the process of design productive. In addition, the distributed environment used by this research is based on the DIM system (Lai and Chang, 2006). By investigating the computational feasibility of a computer-supported system for design decision, this research might provide some insights for increasing the competition edge of Taiwan architectural design.
