Bounding the Price of Anarchy for Mixed Traffic Equilibrium Assignment with Stochastic Demands and Road Pricing
摘要
This paper analyses problems in the price of anarchy (PoA) for a mixed traffic network with stochastic demands. The authors focus on two types of users with distinct path-selection principles, self-interested users (SU) and altruistic users (AU). A variational inequality model for the SU-AU mixed traffic equilibrium assignment with stochastic demands is proposed. The authors develop an upper bound formulation for monomial link travel cost functions by the nonlinear programming approach and analyze the log-normal demand distribution. Additionally, an extended study on the PoA with road pricing for this mixed traffic network is presented and justified. Emphasis is placed on scenarios where road pricing is not included in the total system cost and the other where road pricing is included. The numerical results illustrate that the upper bounds on the PoA are contingent on the degree of the link travel cost function, as well as the maximum/minimum altruism coefficients for both non-road pricing and road pricing implementation. Notably, these bounds are also related to the maximum variation coefficient of the demand and the highest degree of the link travel cost function when the travel demand follows the log-normal distribution.