跳到主要內容

臺灣博碩士論文加值系統

(44.211.117.197) 您好!臺灣時間:2024/05/27 04:58
字體大小: 字級放大   字級縮小   預設字形  
回查詢結果 :::

詳目顯示

我願授權國圖
: 
twitterline
研究生:廖志冠
研究生(外文):Liaw,Jyh-Guann
論文名稱:前例聯想在設計上的應用─以建築平面為例
論文名稱(外文):Applying Case-Based Reasoning to Generate Floor Plans
指導教授:陳珍誠陳珍誠引用關係
指導教授(外文):Chen,Chen-Cheng
學位類別:碩士
校院名稱:國立臺灣大學
系所名稱:土木工程研究所
學門:工程學門
學類:土木工程學類
論文種類:學術論文
論文出版年:1996
畢業學年度:84
語文別:中文
論文頁數:92
中文關鍵詞:前例聯想法知識工程師類神經網路
外文關鍵詞:Case-Based Reasoningknowledge engineerartificial neural networks
相關次數:
  • 被引用被引用:0
  • 點閱點閱:145
  • 評分評分:
  • 下載下載:0
  • 收藏至我的研究室書目清單書目收藏:0
前例聯想法和類神經網路同為人工智慧領域中常用來解決問題的方法之一
,基本上二者間對問題的分析方法不同,前例聯想法是利用前例資料庫來
當做全部的問題空間,並且從中尋求解答。而類神經網路是利用矩陣的運
算,使所有的輸出向量和目標向量間的誤差達到最小,在不斷的調整誤差
過程中,也同時調整各矩陣的相關權值。二者間主要的差異是:前例聯想
法是屬於符號式的方法,而類神經網路則屬於數值式的方法。文中主要是
以建築平面的設計為例子,並且利用前例聯想法及類神經網路來做應用分
析,及其比較。

Exist computational models for floor plan layout generation
rely on generate-and-test method, and require knowledge
engineers as human translators between designers and computers.
To extend the problem solving abilities and to avoid the
knowledge acquisition bottleneck of current design computation
tools, systems allowing the problem-solver to learn from
experience are necessary. One possibility of knowledge
acquisition is to learn from design cases supplied by
designers. Case-based reasoning provides a model for applying
past experience directly to new problems, which emphasizes on
memory retrieval rather than on computing solutions. The
retrieved case may either match the current situation exactly
or need modification. This leads to the fundamental assumptions
of this research that applying case-based reasoning and
artificial neural networks on knowledge-based design systems.
In order to investigate the assumptions, this research proposes
computational models for layout design which draws on
traditional symbolic approach of case-based reasoning in
machine learning together with a numerical approach of
artificial neural networks. As well as, computer systems are
developed to demonstrate different approaches of proposed
model. The results show that applying case-based reasoning and
artificial neural networks for building knowledge-based design
systems are feasible approaches.

QRCODE
 
 
 
 
 
                                                                                                                                                                                                                                                                                                                                                                                                               
第一頁 上一頁 下一頁 最後一頁 top