摘要
在假设各起始点的交通发生量和吸引量确定的情况下,研究不确定需求及不确定建设成本下的连续交通网络设计问题.利用鲁棒优化方法建立了一个不确定双层网络模型,设计了基于Frank-Wolfe法的情景局部搜索法.双层不确定模型中下层模型是需求属于一个简单有界闭集的用户均衡问题,上层模型是建设成本不确定的系统总阻抗最小的优化问题.通过数值算例分析,得到鲁棒优化下的不确定网络设计模型稳定性优于确定性设计模型的结论.
The trip generation of every origin and trip attraction of every destination are assumed deterministic, then the continuous transportation network design problem with demand and cost un-certainty is studied. A bi-level uncertain model is built with robust optimizations. The scenario-based local search algorithm based on the Frank-Wolfe method is designed. The lower-level problem is a user equilibrium problem with the demands belong to a bounded closed set. The upper-level problem is a minimum problem of the impedance of system with construction cost uncertainty. The result of the numerical experiment shows that the robust model with uncertain data has the advan- tage of stability over the deterministic solutions.
出处
《内蒙古大学学报（自然科学版）》
CAS
CSCD
北大核心
2014年第1期91-97,共7页
Journal of Inner Mongolia University：Natural Science Edition
基金
国家自然科学基金(71061010
71262008)
教育部留学回国人员科研启动基金
内蒙古大学高层次人才引进科研启动项目(210221)
内蒙古自治区高等学校"青年科技英才计划"项目(NJYT-13-B02)
内蒙古自治区自然科学基金项目(2010MS0705)
关键词
需求不确定
建设成本不确定
双层网络设计
鲁棒优化
情景局部搜索法
demand uncertainty
cost uncertaintyscenario-based local search algorithm
bi-level network design
robust optimizations
