重庆理工大学学报2025，Vol.39Issue(19)：13-20,8.DOI:10.3969/j.issn.1674-8425(z).2025.10.002
混合高斯分布的随机交通网络最可靠路径改进交替乘子方向法
Improved alternating direction method of multipliers for the most reliable path in stochastic traffic networks with the Gaussian mixture distribution
摘要
Abstract
To simulate the route selection behavior of vehicles under different traffic conditions,a mathematical model for the most reliable path in a stochastic traffic network with the Gaussian mixture distribution is built and solved.The Gaussian mixture distribution is employed to fit the link travel times in the random network,and the augmented Lagrangian relaxation method is employed to build the dual problem.It is then solved using the block coordinate descent method inherent in the alternating direction method of multipliers(ADMM),which provides upper and lower bounds for the optimal value.Subsequent iterative approximations yield an approximate optimal solution for the initial problem.Numerical experiments are conducted on the Sioux Fall network and the Chicago Sketch network.Results indicate the Gaussian mixture distribution accurately fits the link travel times under different traffic conditions.The most reliable path between the start and the destination varies under different traffic conditions.The ADMM algorithm proves effective in finding the most reliable path under the Gaussian mixture distribution,demonstrating superior accuracy and convergence compared to other methods.关键词
交通网络/随机网络/混合高斯分布/最可靠路径/交替乘子方向法/块坐标下降Key words
traffic network/stochastic network/the Gaussian mixture distribution/the most reliable path/ADMM/block coordinate descent分类
交通运输引用本文复制引用
潘义勇,刘宇,曹天宇..混合高斯分布的随机交通网络最可靠路径改进交替乘子方向法[J].重庆理工大学学报,2025,39(19):13-20,8.基金项目
国家自然科学基金项目(51508280) （51508280）
