Study Of Optimal Path In Traffic Network With Random Link Travel Times Based On Real-time Traffic Information

The optimal path is the key problem of route guidance system in intelligent transportation system.The optimal path in traffic network which random link travel times based on real-time traffic information is the basis of vehicle navigation system.Combined with the real-time traffic information,this model can select and adjust the optimal path in time,which not only has the effect of reducing the travel time of travelers,the effective use of road resources,but also has an important significance to alleviate road traffic congestion and equilibrium traffic flow.The paper put the optimal path of transportation network as the research object,which was based on traffic flow theory,model theory and so on.This paper's specific content is as follows:(1)Based on the previous researches o f road resistance function,considering the property of urban roads,it was proposed that the road resistance was the sum of the road resistance and node resistance.According to the division of traffic state,this paper established 4 kinds of the road resi stance function which corresponded to smooth,amble,congestion and severe congestion.Then,using the influence factors,such as intersection interval,non-motor vehicle interference,pedestrian interference,lane width,and so on,to modify and optimize resistance function.(2)Manifestation of the nodes,edges,weights and adjacency list which have small storage of traffic network were determined on the basis of graph theory.The probability distribution image of travel time which obtained by vehicle-following method proved that the weight is random.Based on that,the paper proposed that neglecting the traveler's choice behavior in the static stochastic traffic network during the process of travel,while ignoring the change of traffic state in traffic network,the optimal path problem was transformed into a nonlinear integer programming problem by using the network equilibrium principle.Then taking the minimum travel time as the optimal goal to establish the path function model.At the same time,prior vehicle path suggestion was prop osed to solve the model.(3)Comparison between adaptive decision and prior decision,adaptive decision which means that continuously optimize the travel path is better than prior decision during the process of driving.This paper expounded the advantages of a daptive decision,defined the adaptive search range,and integrated real-time traffic information into the objective function.Using the network equilibrium principle to construct equilibrium relation of the current node,constructing the optimal path model based on real-time traffic conditions,then through the cycle of searching real-time information and solving the optimal path,the state chains of path were obtained and the optimal path can be gotten.At the same time,adaptive vehicle path suggestion was proposed to solve the model.(4)The prior optimal path model and the BPR function model were tested by Sioux Falls(SF)network respectively in two different hypotheses.Compared with the result obtained by traditional model,the difference of prior optimal path model 's selected path was not obvious.Adaptive optimal path model based on real-time traffic information used a topological map which transformed from a regional road network in Nanjing,and set up two sets of assumptions to conduct numerical experiments.The results of adaptive optimal path model were compared with the results of a prior optimal path model and to summary the applicability of the two models.At the same time,the results proved that the adaptive optimal pat h model was more suitable for the application in real life.By the research of this paper,the road resistance function model,a prior optimal path model and an adaptive optimal path model were constructed in the static stochastic network environment.Usin g the algorithm,MATLAB programming to solve the models,the feasibility of the model was verified and the applicability of the model was analyzed through numerical experiments.This paper provides a theoretical basis for the research of intelligent route guidance system.