基于ANP与麻雀搜索算法优化BP神经网络的装配式建筑进度风险评价

Schedule Risk Evaluation of Prefabricated Building Based on ANP and SSA-BP

进行进度风险评价是装配式建筑风险管理的重点,也是提高装配式建筑工期效益的关键。从设计生产、运输供应、施工管理及环境4个方面识别影响因素并进行风险评价指标体系构建,在此基础上采用网络层次分析法(ANP)构建指标关系网络并进行权重计算。同时针对传统BP神经网络评估精度低的问题,引入麻雀搜索算法(SSA)优化BP神经网络,建立起基于ANP与SSA-BP神经网络的评价模型实现对装配式建筑进度风险的评价研究,通过与ANP-BP神经网络评价模型的对比验证了该模型的准确性与稳定性。

Schedule risk evaluation is the key to risk management of prefabricated buildings and improvement of the prefabricated building construction efficiency.This paper identifies the influencing factors and constructs a risk evaluation index system from design and production,transport and supply,construction management,and environment.In this way,ANP is used to construct the index relationship network and calculate their weight.Meanwhile,SSA is introduced for the problem of low accuracy of traditional BP neural evaluation.The evaluation model is established based on an ANP and SSA-BP to evaluate the schedule risk of prefabricated buildings.Compared with the ANP-BP evaluation model,the proposed model has more accuracy and stability.