主要讨论基于用户平衡原则的交通网络优化问题.这些问题大致上可以分为二大类:一类是涉及到确定性用户平衡原则;另一类是考虑随机性用户平衡原则.众所周知,运筹学中的双层规划模型能够完美地刻划这些问题,但是所建立的双层优化模型往往属于不可微优化问题的范畴,这就给设计有效的算法带来了很大困难.此文首先从模型和算法的角度总结了有关这类问题已有的研究成果,接着介绍有关这方面的最新的研究进展,即如何把用户基于平衡原则下的交通网络优化问题的双层规划模型统一地转换为一个连续可微的单层最优化问题,并设计统一的算法.作为统一的算法方面的研究,我们可以看到增广的拉格朗日方法可以用来解上述的第一类问题,而基于灵敏度的分析的序列二次规划方法完全有能力解上述的第二类问题. Mainly discusses the traffic network optimization problem based on the principle of user balance.These problems can be roughly divided into two categories: one is related to the principle of deterministic user balance and the other is the principle of stochastic user balance.As we all know, operational research The bilevel programming model perfectly describes these problems, but the established two-tier optimization model often falls into the category of non-differentiable optimization problems, which brings great difficulties to the design of effective algorithms.This paper starts from the model and From the perspective of the algorithm, the existing research results on this kind of problems are summarized, and the latest research progress in this area is introduced. That is, how to uniformly convert the user’s bi-level programming model of traffic network optimization based on the principle of equilibrium into a continuous We can see that the augmented Lagrangian method can be used to solve the first type of problem, while the sensitivity-based The analysis of the sequence quadratic programming method is fully capable of solving the second type of problem.