沈阳建筑大学学报:社会科学版
瀋暘建築大學學報:社會科學版
침양건축대학학보:사회과학판
Journal of Shenyang Jianzhu University:social Science
2012年
2期
152-156
,共5页
齐宝库%牛驰野%赵景明%邵帅
齊寶庫%牛馳野%趙景明%邵帥
제보고%우치야%조경명%소수
绿色建筑评价%绿色建筑评价指标集%BP神经网络%梯度下降法%阀值
綠色建築評價%綠色建築評價指標集%BP神經網絡%梯度下降法%閥值
록색건축평개%록색건축평개지표집%BP신경망락%제도하강법%벌치
green building evaluation%evaluation index collection of green building%BP-neural network%gradient descent algorithm%threshold
将人工神经元网络应用于绿色建筑评价中,建立了采用LM算法的三层BP神经元网络绿色建筑评价模型,并参考国内外对绿色建筑评价的研究,建立绿色建筑评价指标集,通过专家打分获得训练样本,用trainlm训练函数训练网络,验证了绿色建筑评价结果与实际的一致性。最后,通过已训练好的网络,对多个绿色建筑方案进行评价选优,验证了该模型在绿色建筑评价中的实用性与可靠性。
將人工神經元網絡應用于綠色建築評價中,建立瞭採用LM算法的三層BP神經元網絡綠色建築評價模型,併參攷國內外對綠色建築評價的研究,建立綠色建築評價指標集,通過專傢打分穫得訓練樣本,用trainlm訓練函數訓練網絡,驗證瞭綠色建築評價結果與實際的一緻性。最後,通過已訓練好的網絡,對多箇綠色建築方案進行評價選優,驗證瞭該模型在綠色建築評價中的實用性與可靠性。
장인공신경원망락응용우록색건축평개중,건립료채용LM산법적삼층BP신경원망락록색건축평개모형,병삼고국내외대록색건축평개적연구,건립록색건축평개지표집,통과전가타분획득훈련양본,용trainlm훈련함수훈련망락,험증료록색건축평개결과여실제적일치성。최후,통과이훈련호적망락,대다개록색건축방안진행평개선우,험증료해모형재록색건축평개중적실용성여가고성。
The paper applies the artificial neutral to green building evaluation. The model of evaluation for green building based on three-layer BP-network is established by using LM algorithm. By domestic and foreign research for green building evaluation, the evaluation index collection of green building is selected. The network is trained by trainlm function with training samples that have been obtained by expert. The re- sult is consistent with the actual work by the network and the applicability of the evaluation model is illus- trated by examples. Finally, this paper chooses the best one of several green building schemes by using trained network. The practicability and reliability of the model can be verified in this process.