摘要
在地铁建设过程中,车站作为其不可或缺的组成部分,其施工工期对整个项目的进度控制和施工期间资源的合理配置具有重要作用。因此,实现地铁车站施工工期的精准预测至关重要。文章在深入分析暗挖法地铁车站施工工期影响因素的基础上,基于Spearman相关性分析筛选出18个影响因素,并采用SPSO算法优化BP神经网络,构建基于SPSO-BP的工期预测模型。实验结果表明,相比于传统的BP神经网络、PSO-BP模型以及未经特征筛选的SPSO-B P模型,SPSO-BP模型在小样本的施工工期预测上具有更高的预测精度和效率。
In the process of urban rail transit construction,the construction periods of the indispensable metro stations are of crucial significance for the progress control of the entire project and the efficient allocation of resources during construction.Therefore,it is important to accurately predict the construction period of metro stations.Based on the in-depth analysis of the influencing factors affecting the construction period of mined metro stations,this paper screens out 18 influencing factors based on Spearman correlation analysis,and uses improved standard particle swarm optimization(SPSO)to optimize BP neural network to construct a construction period prediction model based on SPSO-BP.Experimental results show that compared with the traditional BP neural network,PSO-BP model and SPSO-BP model without feature screening,the SPSO-BP model has higher prediction accuracy and efficiency in the prediction of construction period with small samples.
出处
《现代城市轨道交通》
2023年第8期30-36,共7页
Modern Urban Transit
基金
国家自然科学基金(71861022)。
