| China is in a critical period of rapid development in social economy and urbanization,with continuous expansion of urban scale,rapid agglomeration of population and rapid growth of car ownership.When the imbalance of supply and demand occurs in the motorized travelling of urban traffic system,there will be many problems represented by traffic congestion and pollution of exhaust emission.As a key means to realize the dynamic balance between traffic supply and demand,traffic management strategy(TMS)on vehicles has been widely used in various urban traffic systems.In view of the high complexity of urban traffic system,the effectiveness of a TMS is often difficult to evaluate in advance,and the decision scheme is often adjusted according to the feedback of the actual implementation effect.However,trial and error of management decision in the actual traffic system will inevitably lead to man-made traffic disorder.Therefore,constructing the urban virtual traffic system and realizing the accurate evaluation of the implementation effect before implementation will become the prerequisite for realizing the decision support of traffic management in the construction of intelligent traffic.The TMS aims to alleviate urban traffic congestion and control exhaust emissions by controlling the quantity and distribution of various types of motor vehicles in the road network,and to guarantee the operation quality of the traffic network at the same time.In the decisionmaking process of general management strategy,the effectiveness estimation and strategy optimization of the scheme are essential.In the existing TMS related research,the characteristics analysis of travel demand structure changing under temporary and regular implementation of the TMS is insufficient,thus,it is difficult to effectively estimate the shortterm and long-term traffic equilibrium states.In the process of management scheme optimization,the implementation purpose is usually single,which is difficult to meet the practical objective of strategy optimization under diversified implementation purposes.The collaborative optimization design of joint TMS is still difficult to implement and inconvenient to be applied in practice.Vehicle traffic management for future traffic environment with intelligent network connection and autonomous driving technique has not been involved,and the research field is still blank.Therefore,in view of the deficiencies and defects of these studies,the research on TMS and its optimization techniques will make an important contribution to the decision support in the urban virtual traffic system.Accordingly,based on the topic "information service and decision evaluation technology in the urban virtual transportation system" of the national key research and development plan(2016YFE0206800)named "construction technology of ’One Belt And One Road’ urban intelligent transportation system under complex environment",this paper carries out the research on TMS analysis and related optimization techniques.On the basis of the literature review and the analysis of basic method for equilibrium assignment,research work in this paper is firstly conducted on the short-term and long-term equilibrium assignment under TMS,and then focus on the joint optimization model development.Thereafter,joint TMS of traffic restriction based on license plate considering carpooling exemption and that considering congestion charging are introduced and related optimization models are developed.Furthermore,combined with the future intelligent network traffic environment,a novel TMS in the mixed traffic flow with connected autonomous vehicles(CAVs)and human-driven vehicles(HVs)is proposed and its optimization model is established.Finally,based on the simulation software "TranStar",the decision-making support function is developed and case study of TMS is then conducted.The main research work of this paper includes the following five aspects:(1)Study on the short-term traffic equilibrium under TMSFirstly,the theoretical basis of traffic equilibrium assignment is sorted out,and then the short-term influence of first-time and temporary TMS on urban traffic equilibrium is analyzed.After the implementation of the short-term TMS,based on the principle of minimum travel cost,the route choice of the regulated travel mode is re-analyzed,and the travel cost of all travel modes is re-evaluated.Accordingly,by integrating the regional location of TMS and the detour distance of the restricted travel mode,the mode shifting rate of the restricted traffic demand is calculated.Then,the short-term equilibrium assignment model under TMS is established,and the corresponding solution algorithm is proposed.The Sioux Falls network is used to verify the proposed model by comparing it with the existing model without considering mode shifting rate.Finally,the sensitivity analysis of the basic parameters used in the model is carried out to explore the influence of different parameter settings on the mode shifting rate.(2)Study on the long-term traffic equilibrium under TMSFirst of all,under the assumption that the network’s potential travel willingness is stable,a general elastic demand model for generalized travel costs related to actual total travel demand is put forward under the long-term traffic management.Then,combining analysis of joint choice of travel mode and path based on Nested-Logit model,a long-term equilibrium assignment model for general elastic demand is proposed,and an improved directional search algorithm is introduced to solve this model.Second,as for traffic restriction management strategy based on license plate number,special consideration should be given to the impact of TMS on private car ownership.Under the assumption that the actual total travel demand is stable,the structure of travel demand considering the change of private car ownership is analyzed,and the quantitative relationship between the change and reorganization of travel demand structure is analyzed.Combining the analysis on travel cost,the elastic demand model is established considering private-car ownership changes,and a solution algorithm integrating simplified diagonalization algorithm,Gauss Seidel decomposition and MSA algorithm are designed.Finally,a numerical experiment is carried out under Sioux-Falls network to verify the feasibility and effectiveness of proposed model and algorithm.(3)Study on the optimization techniques of TMSFirstly,with the determination of optimization objectives for TMS,a implementing purpose oriented bi-level programming model is proposed to achieve the collaborative optimization of traffic management area and proportion,of which the upper-level model focus on the maximization of travel convenience,with comprehensive considering the total travel time of the system which reflects network’s exhaust emission characteristics and the total overload of the network which reflects network’s traffic congestion characteristics;the lowerlevel model realize the long-term equilibrium assignment of multi-type reorganization traffic demand under TMS.Thereafter,a genetic algorithm based on graph theory is introduced to solve proposed bi-level programming model.Secondly,a joint TMS of traffic restriction based on license plate number and carpooling is proposed,that is,a certain proportion of restricted vehicles are allowed to travel in the restricted area by carpooling,and a collaborative optimization model of this joint strategy is constructed.Then,another joint TMS of traffic restriction based on license plate number and congestion charging is introduced,that is,the restricted vehicles are allowed to travel in the restricted area by paying the congestion fee,and the collaborative optimization model of this joint strategy is built.Finally,the feasibility and effectiveness of the proposed optimization model are verified by analyzing the multi-scene optimization scheme under Sioux-Falls network.(4)Research on TMS in the prospective intelligent network traffic environmentFirst of all,based on existing research findings in the field of intelligent network traffic environment is summarized and analyzed,the capacity model considering HVs flow mixed with CAVs is given,and the travel time of mixed flow section is analyzed in combination with the equivalent coefficient.Secondly,the hybrid traffic equilibrium model is built and the Manifold sub-optimization algorithm is introduced to solve the strong stationary solution of the optimal/worst equilibrium.Then,with the controllable advantages of CAVs,the UE-SO hybrid equilibrium model of mixed flow is proposed.Accordingly,a novel TMS for mixed traffic flow without considering lane management is proposed,and the optimization model of differentiated TMS for CAVs is established,which is solved by sensitivity analysis and alternating direction method of multipliers.Furthermore,the traffic equilibrium model for mixed flow joint TMS considering lane management is established,and the optimization model of mixed-flow lane management scheme is developed and the corresponding solution algorithm is designed.Finally,based on Sioux-Falls network and Nguyen-Dupuis network,numerical experiments are carried out to verify the rationality of the proposed joint TMS for mixed flow,the validity of the proposed models and the feasibility of the proposed algorithms.(5)Case study of TMS based on the urban virtual traffic systemFirst of all,this paper introduces the test platform of urban virtual traffic system,that is"TranStar",sorts out the software modules involved in the development of decision support function of urban TMS,and determines the development framework based on TranStar software.Secondly,the software implementation of TMS decision analysis function in urban virtual traffic system is gradually completed from four aspects,namely,topological refinement of basic traffic network,information perception of infrastructure,semantic description of management strategy and integrated analysis of simulation operation.Then,the virtual simulation database of Chongqing is constructed,and OD matrix and network cross-section flow are extracted through Chongqing’s radio frequency identification devices data mining,and the simulation model parameters are calibrated.Furthermore,based on the humancomputer interaction graphics editing function of TranStar software,the simulation input of TMS is completed,and the "one-button" simulation operation integration analysis module is configured to realize the decision-making evaluation of TMS.Finally,through the comparison analysis of comprehensive evaluation indexes before and after the implementation of tested schemes,the effectiveness and practicability of the developed traffic management decision analysis function is verified.,... |
PDF Full Text Request