Price of Anarchy in Transportation Networks with Heterogeneous Agents

This article proposes a model to analyze the transporttion networks with heterogeneous agents, who contribute to congestion differently. This is the generalization of homogeneous model. It turns out that the optimum from social planning is high dependent of the distribution of the driving technique possessed by agents in the networks, both for its mean value and variance. However, the optimum from selfish view is only dependent of the mean value of driving technique. Price of anarchy analysis displays that there is more great area to operate profitably for social planning under the heterogeneity, comparative to that under the homogeneity. Because of difficulty of technique recognization for the social planner, as a signal to coordinate selfish agents to get social optimum, charge must be changed with the distribution of the driving technique possessed by agents in the networks.