基于改进AHP-BP的深基坑施工风险评估

Risk Assessment of Deep Foundation Pit Construction Based on Improved AHP-BP

由于深基坑施工过程风险影响因素较多,传统的专家打分法和层次分析法计算得到的指标权重存在较多的无效指标和较强的主观性。为了对深基坑施工风险进行准确评估,提出DEA-AHP和BP神经网络的深基坑施工风险评估模型。利用WBS-RBS对深基坑施工项目中潜在的风险进行识别,得到风险指标。采用主成分分析法和数据包络分析法实现评价指标的二次筛选,简化判断矩阵,降低冗余指标对管理者的干扰;再利用层次分析法对余下指标进行赋权,并将指标权重作为学习样本,训练BP神经网络,实现风险等级预测。通过实证分析,能准确得到风险等级,为相关研究提供参考。

Because there are many risk factors in the construction process of a deep foundation pit,the index weights calculated by the traditional expert scoring method and the analytic hierarchy process have many invalid indexes and strong subjectivity.To accurately assess the construction risk of a deep foundation pit,the DEA-AHP and BP neural network risk assessment model of deep foundation pit construction is proposed.Firstly,WBS-RBS is used to identify the potential risks in deep foundation pit construction projects,and risk indicators are obtained.Secondly,principal component analysis and data envelopment analysis are used to realize the secondary screening of evaluation indicators,simplify the judgment matrix,and reduce the interference of redundant indicators on managers.Thirdly,the remaining indicators are weighted using the analytic hierarchy process.Finally,the weights of indicators are used as learning samples to train the BP neural network to achieve risk level prediction.Through empirical analysis,the risk level can be accurately obtained,which provides references for relevant research.