基于贝叶斯网络的一种新型的结构选型方法研究(Ⅱ)

Research on a new approach to structure system selection of high-rise buildings based on Bayesian networks (Ⅱ)

最佳高层建筑的结构型式的选取与确定是结构工程初步设计的重要内容,然而,结构型式的选择受到多种因素的影响与制约,是一个具有高度白噪声、非线性的复杂问题。针对影响结构型式选择的各种因素,首先分别构建了影响因素矩阵向量与结构型式矩阵向量,引入灰色关联理论对各影响因素的重要程度进行了从定性到定量的分析。根据获取的影响因子的重要性程度,确定了简化的关键影响因素矩阵。在此基础上,引入贝叶斯网络理论,建立结构选型的贝叶斯网络预测模型。以3种情况为实例,分别验证了该模型的有效性与计算精度。试验结果表明,与其他方法相比,该模型简单实用,且具有较高的预测精度,为结构选型提供了一种新的预测方法。

The optimum structure system selection is the important task in the preliminary design of structural engineering.However,the selection of structure system was influenced by many factors,and is a highly white noise and non-linear complex problems.Considered these various factors,first of all the influencing factors vector matrix and structure system vector matrix were constructed.The grey relational theory was introduced to analysis the importance of influencing factors from qualitative to quantitative.According to obtained the importance degree of the influencing factors,the key influencing factors matrix was constructed.On this basis,the Bayesian networks theory was introduced to set up the structure system selection prediction model.As an example to the three situations,the validity of the model and calculation accuracy had been verified.The calculated results showed that compared with other methods,the model is simple and practical,and had a higher prediction accuracy for the structure of selection,also provided a new prediction method.