Day-to-day Traffic Swapping Behavior Models And Intervention Strategy Considering Travelers’ Micro-preferences

An urban transport network may be always in the non-equilibrium states due to ceaseless internal and external disturbances.Therefore,it is unsuitable to analyze it by the traditional equilibrium theory However,the current non-equilibrium network traffic modeling methods do not lay adequate consideration on travelers’ micro-preferences,and the existing intervention strategies are also unable to fullfill the management demand in dynamical traffic conditions.In view of these,this dissertation deploys a variety of disciplinary theories and methods,and establishes two day-to-day dynamics models in which travelers’ macroscopic property and micro-preferences in traffic swapping behavior are fully considered.In addition,a dynamic dayparting pricing strategy is suggested to intervene the network traffic flow evolution in more efficient way.It is expected that this strategy can provide auxiliary support for realistic traffic management and decision-making.The main contributions and works of this dissertation are summarized as follows.(1)This dissertation argues that it is not only the macroscopic criterion(i.e.,rational behavior adjustment criterion)but also the marginal cost preference as well as the other two suggested micro-preferences(i.e.,the simplicity and the proximity preference)that should be incorporated in day-to-day traffic swapping behavior modeling.Along this line and by adding the departure-time dimension onto an intended route swapping model,a day-to-day departure time-route swapping process(DTRSP)in a pairwise fashion is established.DTRSP is able to make the macroscopic criteria and three micro-preferences compatible,while avoiding the over-swapping problem.Several critical properties for DTRSP are verified theoretically.Numerical simulation is employed to explore the stability of DTRSP and the influence of micro-preferences on swapping behavior.It is observed from the numerical results that three micro-preferences show tangible effects on the traffic evolution of DTRSP,especially the simplicity preference.(2)By following the modeling logic of DTRSP,a day-to-day departure time-route swapping process(MDTRSP)in a min-cost-pursued swapping manner is proposed.MDTRSP is capable of making the macroscopic criteria and three micro-preferences compatible,and possesses superior behavior intuition.It is suggested from the numerical sensitivity and comparative analyses that i)the stability of MDTRSP is similar to that of DTRSP,and the simplicity preference shows dominant effect,and ii)the swapping manner can influence the network flow evolution,but it does not affect the final evolution result.(3)Considering the time-varying feature of traffic conditions within a day,a day-to-day dayparting pricing strategy is proposed and the corresponding dynamic road pricing model is then constructed.The pricing model takes system optimization as its objective,and captures the day-to-day evolution via DTRSP.Considering that the objective function is non-differentiable and does not possess an explicit and closed-form expression,this dissertation employs a constrained compass search algorithm which needs no any derivative information to solve it.Numerical simulations are conducted to characterize the proposed strategy and model.The results suggest that i)the current compass search method is effective to solve the model,and ii)the proposed day-to-day dayparting pricing strategy is able to decrease the total evolution travel time,and increasing improvement is derived if the scope to implement the strategy is amplified.