混合网络均衡模型的求解算法研究

发布时间：2018-12-13 04:09

【摘要】：为了解决不同属性用户路径选择相互影响的混合交通均衡分配问题,利用变分不等式描述了混合网络中用户平衡UE、系统最优SO、古诺-纳什均衡CN这3类用户的路径选择行为。分别采用基于路段的对角化算法和基于路径的双重投影算法对混合网络均衡模型进行求解,以Sioux Falls网络为例,从计算效率、收敛精度2个方面对其进行对比分析。分析结果表明:对角化算法前期收敛较快,但拖尾严重,很难通过多次迭代达到较高的精度;双重投影算法即使经过上千次迭代,精度仍可提高。在多种混合比例下,SO、CN用户控制流量比越大,系统总阻抗越小,而SO用户流量比低于0.5时,系统总阻抗下降很快,随着流量比增大,下降速度随之变缓。
[Abstract]:In order to solve the problem of hybrid traffic equilibrium assignment in which user path selection with different attributes affects each other, the optimal SO, of user balanced UE, system in hybrid networks is described by using variational inequalities. The path selection behavior of the three classes of users, CN, in the Gunot-Nash equilibrium. The diagonalization algorithm based on road section and the dual projection algorithm based on path are used to solve the equalization model of hybrid network respectively. The Sioux Falls network is taken as an example to compare and analyze it from two aspects of calculation efficiency and convergence precision. The results show that the diagonalization algorithm converges fast in the early stage, but the trailing is very serious, so it is difficult to achieve higher precision through multiple iterations, and the accuracy of the double projection algorithm can be improved even after thousands of iterations. The larger the SO,CN user control flow ratio is, the smaller the total impedance of the system is, but when the SO user flow ratio is lower than 0.5, the total impedance of the system decreases rapidly, and with the increase of the flow ratio, the decrease speed becomes slower.
【作者单位】：北京交通大学交通运输学院;
【基金】：国家自然科学基金项目(60870014) 高等学校博士学科点专项科研基金项目(20130009110001)
【分类号】：U491.123

【参考文献】