【摘要】 文章针对传统的建筑施工安全性分析方法存在的不足,提出了一种基于梯形模糊贝叶斯网络(Bayesian network,BN)的安全性分析方法。该方法将广义区间梯形模糊数理论引入贝叶斯网络安全性分析中,避免了安全性影响因素多状态的精确概率难以获得的状况,用广义区间梯形模糊数描述专家的模糊语言值,通过解模糊化、均值化和归一化得到不同状态发生概率的精确值。将其代入贝叶斯网络中,计算叶节点不同状态的发生概率,进而计算各根节点的后验概率及安全性重要度。该方法应用到建筑施工的安全性分析实例中,实验结果表明能提高贝叶斯网络处理不确定性安全问题的准确性和应用范围。更多还原

【Abstract】 For the insufficiencies of traditional building construction safety analysis methods,a safety analysis method based on Bayesian network(BN)using trapezoidal fuzzy theory is proposed.This method introduces generalized interval-valued trapezoidal fuzzy number theory to BN safety analysis,avoiding the situation that exact probability is difficult to obtain because of multi-state in factors influencing safety.The generalized interval-valued trapezoidal fuzzy numbers are used to describe expert fuzzy linguistic evaluation opinion,and the precision value of the probability of different state can be obtained by means of equalization,ambiguity resolution and normalization.Next,the precision value is taken into BN,the probability of leaf nodes in different state is calculated,and then the posterior probability of each root node and the importance degree of safety are calculated.The method is applied to the instance analysis of construction safety,and the result of the experiment shows that the method can improve the accuracy of BN in handling the uncertain safety issues and extend the application range.更多还原