建築機能單元的配置是一個重要的設計工作。在複雜的建築物中，好的配
置設計在金錢與時間上的節省變得更為明顯。本研究將建築機能單元的配置公
式化為一個組合最佳化問題，並採用融合了許多模擬退火與Hopfield 神經網路
的特徵的退火神經網路來求解，並以一個具有28 個機能單元的綜合醫院診療
大樓作個案研究，此外，也詳細探討退火神經網路的參數對解答品質的影響。
研究結果顯示：(1) 退火神經網路對於解決建築機能單元的配置問題相當具有
效率；(2) 無論參數組合為何，以30 個隨機的初始狀態所得的合法解中之最佳
解差異很小

Architecture layout design is an important design activity. The impact
of good layout practices on money and time saving becomes more obvious
in complex architecture. In this study, the layout problem was formulated
as a combinatorial optimization problem. An annealed neural network
model, which merges many features of simulated annealing and Hopfield
neural networks, was employed to solve the problem. A case study of a
hospital building with 28 facilities was employed to illustrate the practical applications. In addition, the effects of various parameters in annealed neural network were examined. Research reported in this paper leads to the following conclusions. (1) An annealed neural network model is rather efficient in solving the architecture layout problem. (2) Whatever the combinations of the parameters are, the difference of quality between the optimum solutions among 30 feasible solutions gotten from a random initial state is rather small.