|
发表于 2005-5-7 13:25:22
|
显示全部楼层
城市交通网络均衡分配的模型与算法,学术报告讲稿,供参考--黄海军
附件为NE Stier-Moses的博士论文《Selfish versus coordinated routing in network games》,2004,导师为Andreas S. Schulz。
摘要:
A common assumption in network optimization models is that a central authority con-trols the whole system. However, in some applications there are independent users, and assuming that they will follow directions given by an authority is not realistic. Individ-uals will only accept directives if they are in their own interest or if there are incentives that encourage them to do so. Actually, it would be much easier to let users make their own decisions hoping that the outcome will be close to the authority';s goals. Our main contribution is to show that, in static networks subject to congestion, users'; selfish decisions drive the system close to optimality with respect to various common objectives.This connection to individual decision making proves fruitful; not only does it provide us with insights and additional understanding of network problems, but it also allows us to design approximation algorithms for computationally di±cult problems.
More specialcally, the con°icting objectives of the users prompt the de¯nition of a network game in which they minimize their own latencies. We show that the so-called price of anarchy is small in a quite general setting. Namely, for networks with side constraints and non-convex, non-di®erentiable, and even discontinuous latency functions, we show that although an arbitrary equilibrium need not be e±cient, the total latency of the best equilibrium is close to that of an optimal solution.
In addition, when the measure of the solution quality is the maximum latency, equilibria in networks without constraints are also near-optimal. We provide the ¯rst analysis of the problem of minimizing that objective in static networks with congestion. As this problem is NP-hard, computing an equilibrium represents a constant-factor approximation algorithm. In some situations, the network authority might still want to do better than in equilibrium. We propose to use a solution that minimizes the total latency, subject to constraints designed to improve the solution';s fairness. For several real-world instances, we compute tra±c assignments of notably smaller total latency than an equilibrium, yet of similar fairness. Furthermore, we provide theoretical results that explain the conclusions derived from the computational study.
主要工作:
One of our main contributions is an extension of the results of Roughgarden and Tardos to networks with side constraints and to latency functions that are non-convex,non-di®erentiable, and even discontinuous. Side constraints are especially useful in traffic networks. For example, they can be used to model arc capacities, which has been recognized as an important means to provide a more accurate description of tra±c °ows.
Moreover, real-world transportation models often include such constraints, not only to prevent arcs from carrying arbitrarily large °ows, but also as a way to calibrate the network model and bring its solution into agreement with anticipated results. In this more general model, multiple user equilibria might exist and an arbitrary user equilib-rium does not need to be close to a system optimum.
Nevertheless, we characterize a particular equilibrium that we call Beckmann user equilibrium, which turns out to be e±cient. De¯ning the coordination ratio of a solution as the ratio between its total travel time and that of the system optimum, we show that the coordination ratio of a Beckmann user equilibrium is not higher than the price of anarchy in the model without side constraints. Beckmann user equilibria arise from an extension of the mathematical programming formulation used to compute user equilibria in the model without side constraints, due to Beckmann, McGuire, and Winsten (1956). Besides having relatively small total travel time, Beckmann user equilibria are attractive because they can be computed e±ciently. This also shows that game-theoretic concepts successfully o®er a way of computing approximate solutions to hard problems. Although there are mul-tiple equilibria and computing the best equilibrium is di±cult, a provably good user equilibrium can be computed in polynomial time.
We have just argued that, in the worst case, equilibria are not too ine±cient. But what happens in the average case? Or at least in real-world examples? It turns out that equilibria are even closer to system optima, which is good news. But not quite. |
|