摘要
对于一个多类别的网络均衡问题,可以通过计算某个辅助问题的容量限制约束相应的乘子向量得到有效收费.本文通过计算拉格朗日函数的鞍点来计算乘子向量.借助于广义拉格朗日函数的稳定性和Uzawa算法非精确解的收敛性,得到鞍点序列的收敛性.其中离散化方法用于最小化广义拉格朗日函数的计算.
For a multiclass network equilibrium problem,the multiplier vectors corresponding to capacity constraints in an auxiliary problem are valid tolls.In this work,we compute the multiplier vectors through computing saddle points of the Lagrangian function.We prove the convergence of saddle points sequence by virtue of the stability of augmented Lagrangian function and the convergence of nonexact solutions in Uzawa algorithm.Discretization method be used to minimize the augmented Lagrangian function.
出处
《应用数学》
CSCD
北大核心
2011年第4期826-832,共7页
Mathematica Applicata
基金
the NSFC(71071035)
the SHFU(KT09-02)
关键词
网络均衡
多类别
乘子
广义拉格朗日函数
离散化
Network equilibrium
Multiclass
Multiplier
Augmented Lagrangian
Discretization