基于CSO-SVR的气膜钢筋混凝土储仓主体结构施工成本预测研究  

Research on Construction Cost Prediction of Main structure of Concrete Dome Using Inflated Forms Based on CSO-SVR

作　　者：李万庆[1] 陈明欣[1] 孟文清[2] 石华旺[2] 

机构地区：[1]河北工程大学管理工程与商学院,河北邯郸056038 [2]河北工程大学土木工程学院,河北邯郸056038

出　　处：《数学的实践与认识》2017年第1期105-111,共7页Mathematics in Practice and Theory

摘　　要：为了较准确的预测气膜钢筋混凝土储仓主体结构施工成本,提出一种鸡群算法(CSO)和支持向量回归机(SVR)结合模型,即CSO-SVR,利用CSO算法对SVR进行寻优得到全局最优解,从而得到具有最佳参数的支持向量回归机模型,通过气膜钢筋混凝土储仓主体结构施工成本数据预测仿真,结果显示:CSO-SVR模型预测精度高于PSO-SVR,GA-SVR,SVR,BPNN等方法,是预测气膜钢筋混凝土储仓主体结构施工成本的有效工具.In order to predict the construction cost of the main structure of concrete dome using inflated forms more accurately,this article makes a detailed analysis of the factors affecting the cost,and presents a model that combines chicken swarm optimization algorithm（CSO） with support vector regression（SVR）,namely CSO-SVR,the CSO algorithm is used to find the optimal solution to the SVR,so as to get the support vector regression machine model with the best parameters.Through the forecast simulation using the construction cost data of main structure of concrete dome using inflated forms,the results demonstrate that the prediction accuracy of CSO-SVR model is higher than PSO-SVR,GA-SVR,SVR,BPNN and so on,it is an effective tool to forecast the construction cost of main structure of concrete dome using inflated forms.

关 键 词：鸡群优化算法 支持向量回归机模型 气膜钢筋混凝土储仓 成本预测 

分 类 号：TU755[建筑科学—建筑技术科学] TU723.3[自动化与计算机技术—控制理论与控制工程] TP18[自动化与计算机技术—控制科学与工程]