摘要
为了设计基于Logit的随机用户均衡问题的高效算法,对传统的随机用户均衡模型熵项进行分解,得到路段型随机用户均衡模型.在分析路段型随机均衡模型及其优化条件的基础上,以数学规划的方法推导其敏感度方程,这相对于变分不等式的方法更加容易接受,同时因为确定型均衡模型是随机用户均衡模型的一种特例,所以此方法同样适用于确定型用户均衡的敏感度分析.以相继平均算法和敏感度矩阵对算例进行求解,两者结果基本吻合.同时对实际遇到的秩亏问题,提出"分段求解"的方法,有效地解决了矩阵无法求逆的现象.
In order to explore an efficient algorithm for the Logit-based stochastic user equilibrium （SUE） problem, the conventional entropy of the SUE model was decomposed to get the link- based SUE model. In this paper, based on analysis of the Link-based SUE model and its optimizing conditions, a mathematical programming method of sensitivity analysis for the model is presented. The method is more likely to be accepted relative to the variational inequality method. Since user equilibrium in a traffic network is an extreme case of SUE, the method can be used for the Wardropian equilibrium also. Numerical examples are solved by the method of successive averages algorithm and the sensitivity matrixes, both results are consistent. The ＂segmented solution＂ method is given to deal with the rank defect, solving the phenomena of the matrixes which can＇t be inversed effectively.
出处
《东南大学学报（自然科学版）》
EI
CAS
CSCD
北大核心
2013年第1期221-225,共5页
Journal of Southeast University：Natural Science Edition
基金
国家自然科学基金资助项目(51078085
51178110)
