求解交通网络平衡问题的并行分裂算法  

Parallel Split Algorithm for Solving Traffic Network Balancing Problem

作　　者：朱雪珂 李欢 ZHU Xueke;LI Huan(Prospect College,Chongqing College of Mobile Communication,Chongqing 401520,China)

机构地区：[1]重庆移通学院远景学院,重庆401520

出　　处：《佳木斯大学学报（自然科学版）》2024年第8期170-172,共3页Journal of Jiamusi University：Natural Science Edition

摘　　要：提出了一种新的并行分裂算法,用于求解能转化为可分离结构的变分不等式的交通网络平衡问题。基于该问题的可分离结构,分裂算法被广泛研究用于求其解。交替方向法是一种经典的求解变分不等式问题的算法,基于交替方向法,考虑当数据维数较大时,并行分裂算法更有效,且并不是所有子变分不等式都能精确求解。本文提出了一种新的非精确并行交替方向法,在适当条件下证明了算法的收敛性,并应用于交通网络平衡问题,最后通过数值结果表明了算法的有效性。A new parallel splitting algorithm is proposed for solving transportation network balance problems with variational inequalities that can be transformed into separable structures.Based on the separable structure of this problem,splitting algorithms have been widely studied to find its solution.The alternating direction method is a classic algorithm for solving variational inequalities.Based on the alternating direction method,it is considered that when the data dimension is large,the parallel splitting algorithm is more effective,and not all sub-variational inequalities can be solved accurately.This paper proposes a new non-exact parallel alternating direction method,proves the convergence of the algorithm under appropriate conditions,and applies it to the traffic network balance problem.Finally,the effectiveness of the algorithm is demonstrated through numerical results.

关 键 词：变分不等式 并行分裂算法 可分离结构 交替方向法 

分 类 号：O224[理学—运筹学与控制论]