Research On Reversible Lane Optimization Based On Hierarchical Programming Model

Aiming at the imbalance of tidal congestion that often occurs in urban transportation networks,especially the imbalance during morning and evening peaks,this paper constructs an effective model.The structure of the optimization model and the optimization algorithm study emphatically,and a variety of variable lane setting schemes are simulated and compared.The optimal design of the variable lane is an optimal scheme of the variable lane setting.After solving the optimal scheme of the variable lane setting,through the external lane changing facilities carry out the lane changing processing to achieve the purpose of relieving tidal traffic.This article builds the model from the total cost of the system as the starting point.Because the flow of traffic reflects the ideas of travelers.In order to meet the validity of the model,constructing a hierarchical planning model.The lower-level model uses the first principle of Wardrop to construct a traffic distribution that satisfies the balanced distribution of households.Due to the spatial imbalance of tidal flow during the morning and evening peaks,the lowerlevel model constructs two traffic distribution models they are the Morning peak distribution model and the Evening peak distribution model.The construction of the upper-level model considers the total cost of driving during the morning and evening rush periods and the cost of traffic management required for the setting of variable lanes.The cost of setting up variable lanes depends on the number of variable lanes and the vehicles on the road.The solution of the model adopts the Frank-Wolfe algorithm of route allocation method in the lower model,and the genetic algorithm in the heuristic algorithm in the upper model.The simulation results of various setting schemes show the effectiveness of the model and algorithm in this paper.Finally,as a further expansion,the design idea of the variable lane setting scheme of hierarchical planning based on the STGCN prediction model is proposed.Using the STGCN model to predict the traffic flow,and a number of basic traffic prediction models compared.The model’s prediction accuracy and training time analyzed,which shows the efficiency of the model.The traffic flow data predicted by the STGCN model can construct dynamic OD demand.The dynamic OD demand combined with the variable lane hierarchical Programming Model produces a variable lane setting division that is more in line with the actual traffic law,and further optimizes the variable lane setting.Rationality.