Day-to-day Evolution Of Traffic Flow Based On Reference Point

Flow in network is one of the theme in the field of transportation research,on which travelers,choice behaviors have significant influence.The evolution process in network is estimating travel time based on their experience and making route choice based on estimated time,which influences flow assignment and affects route choice behavior in the next day in turn.Not only the real travel time,but characteristics of travelers have influence on route choice.Day-to-day evolution models based on reference point are built in this paper.The evolution processes are simulated to study the impact of reference point on the mode choice,route choice and mixed travel behaviors with consideration of autonomous vehicle.Three parts of researches about the influence of reference point on the evolution process are carried out in this paper,which are shown as follows:(1)The day-to-day evolution model based on reference point is established to describe the influence of travelers' feelings on evolution process.It is assumed that travelers can select one of the modes,car,bus or bike-sharing.The real cost of one mode is compared with the average cost in this model to reflect the influence of travelers' feelings.The generalized cost is influenced by many factors,as a result that travelers' mode choice behaviors dynamically change.The appearance of bike-sharing influences the evolution process of mode choice.It is analyzed that the influence of comfort of every mode and the bus frequency on evolution process.(2)The influence of travelers' feelings on the evolution process is studied in this chapter by the evolution model based on reference point,which compares the actual travel time with the reference point.This model reflects that travelers hold different attitudes towards loss and gain by using different value of parameters.First of all,it is discussed about evolution process when the value of reference point dynamically updates,and existence,uniqueness and stability of equilibrium state are analyzed.Then,the evolution processes are discussed under three cases,when the travelers are optimistic,pessimistic and neutral.The evolution results indicate that total travel time is longest when travelers are pessimistic and shortest when travelers are optimistic.(3)The evolution processes under the condition of mixed travel behaviors are studied considering the appearance of autonomous vehicles.There is a hypothesis that route adjustment rule of autonomous vehicles is different from travelers.It is assumed the target of route adjustment of autonomous vehicles is system optimization.While travelers want to minimize travel time of themselves.These different route adjustment principles lead to mixed equilibrium.Based on two different flow adjustment rules,the simulations are carried out to study the influence of the proportion of autonomous vehicles and the different reference points on the evolution process and the total travel time.With increase of number of autonomous vehicle,the total travel time can be decreased.