Revisiting excess demand capacity model and algorithm of traffic network
Urban-road network capacity,which is a key indicator for assessing the balance between urban-traffic supply and demand and a crucial decision-making basis for government investments,re-fers to the maximum traffic flow that a network can accommodate under the constraint of link capaci-ty.Investigations into network capacity enables the prediction of additional demand that can be ac-commodated by a network,thereby assisting managers in formulating practical and effective traffic-planning and management policies.Mathematically,such issues are typically formulated as bi-level optimization models,with the upper-level model maximizing the total traffic demand and the lower-level model modeling traveler path-choice behavior and network-congestion effects.Because of the complexity of solving bi-level optimization models and their impracticality for large-scale networks,most of these bi-level models are limited in practical applications.This study systematically reviews a single-level network capacity model,namely the excess-demand network-capacity model,and pro-poses an efficient solution algorithm.This study introduces the concept of an augmented network to describe the capacity of a road network.The maximum network capacity is obtained by iteratively solving the traffic assignment for each fixed-demand problem.Numerical experimental results show that the excess-demand network-capacity model and its algorithm can accurately predict the maxi-mum traffic-demand capacity of large-scale urban-traffic networks under different cases,thus provid-ing a robust tool for urban-traffic planning and management.In scenarios where the road network is at its capacity limit,certain links whose volumes have reached or approached their capacity may be-come bottleneck links,thus affecting the overall network capacity.This information can be utilized to guide governmental decisions regarding traffic control strategies or road-expansion investments,thereby offering widespread practical applications.
国家科技期刊平台
NSTL
万方数据
