长沙理工大学学报（自然科学版）2025，Vol.22Issue(4)：184-193,10.DOI:10.19951/j.cnki.1672-9331.20220714001
基于改进SSA-BP算法的沙漠公路造价预测研究
Research on cost prediction of desert highway based on improved SSA-BP algorithm
摘要
Abstract
[Purposes]This paper aims to accurately predict the cost of desert highways.[Methods]Taking the desert highway in the Tarim area of Xinjiang as the research object,this paper constructed the cost index system based on the characteristics of the desert highway and adopted the mean impact value method to screen the index with a significant impact on cost.The back propagation(BP)neural network model was optimized by the sparrow search algorithm(SSA)improved by Logistic chaos mapping to improve the optimization ability and prediction accuracy and overcome the problems of small sample size and high problem dimension of desert highways.The prediction models were evaluated using root mean square error and mean absolute percentage error,and the predictive performance of the proposed model was validated by comparative analysis.[Findings]The indexes,including subgrade thickness and special subgrade treatment length,which are more important to the cost of desert highways,are determined,and the cost of the desert highway is predicted,combined with examples.The mean absolute percentage error of the desert highway cost predicted by this model is 4.26％,which is 17.39 percentage points lower than that predicted by the original model.[Conclusions]The proposed model shows excellent prediction performance,which can provide a reference for investors to accurately predict the cost of desert highways in the early stage.关键词
沙漠公路/造价预测/麻雀搜索算法/BP神经网络/Logistic混沌映射Key words
desert highway/cost prediction/sparrow search algorithm/BP neural network/Logistic chaos mapping分类
建筑与水利引用本文复制引用
王首绪,胡杨子彦,赵娜..基于改进SSA-BP算法的沙漠公路造价预测研究[J].长沙理工大学学报（自然科学版）,2025,22(4):184-193,10.基金项目
湖南交通科技进步与创新项目(201330) Hunan Provincial Transportation Science and Technology Progress and Innovation Project(201330) （201330）
