Research On Dynamic Decision-makings Of Urban Transit Passengers

The development of science should be human-oriented.As one of the important representatives of public welfare programmes,urban transit provides the most basic daily traffic services for large numbers of citizens.The development of urban transit could be seen as a significant manifestation for the development of city modernization as well as the improvement of life quality for the common people.With the enrichment and development of urban transit system,its coverage is getting larger and larger while its operation status becomes more various.Although there has been huge improvement in the quality of urban transit services,it has been more complicated and more changeable for passengers to use the system.The dynamic characteristics of travel decisions have become more common and obvious.The patterns of transit passengers are important references for the construction and development of urban transit.So that it is both theoretically and practically important to get better understanding of the decision-making behavior characteristics that passengers have in urban transit trips.The dynamic decision-making problems in travel behaviors have been widely discussed.However,the current studies are mainly based on the economic theory.These methodologies aim to describe the dynamic decision-makings of travelers in the different stages of their trips.However,there has been few researches that are able to explore dynamic decision-making problems from the angle of human,which are the true subject of the problem.This paper conducts the research from the perspective of human minds.By focusing on the mechanisms of human micro psychological activities,experiments and simulations that are based on decision field theory from psychology studies were implemented to explore the dynamic decision-makings of urban transit passengers.The accomplished works and conclusions are as follows.The characteristics of urban multi-mode transit system is analyzed.The decision set is constructed from the transit modes of bus,netztaxi and bus + metro.To avoid the problem of information overload for decision makers caused by the multi-attributes of multi-modes,and to make it more comparable between the transit modes,this paper constructed the decision parameter system with the homogeneous attributes namely: in vehicle time,crowding status,walk distance and travel costs.According to the characteristics of multi-mode transit trips,this paper defines the dynamic decision-makings in two layers.One is the time-varying psychological activities that multi-mode transit passengers have in the decision processes.It is used to describe the micro mechanisms in the formulation of decisions.The other one is the dynamic choices in the stages of travel process.It is used to describe the choice behaviors that multi-mode transit passengers have in different stages of their trips.The dynamic decision making model for urban multi-mode transit passengers is constructed based on the decision field theory that enables to describe both time dynamics and micro representativeness.First,the development process of decision field theory is analyzed and the basic thoughts,structures as well as the key parameters of decision field model are introduced.With the consideration of the heterogeneous in both decision makers and transit modes,the expression dimensions of attention shift probability on attributes are improved,with the combination of comprehensive weighting method,the initial attention degree on attributes,and the weight of scheme attribute attention.The improved model realized the independent distributions of attribute attentions on each scheme by using the two-dimensional attention shift probability matrix,and the attribute attention changed from synchronization to asynchronization It could avoid the normalized problem that traditional model has in the description of human mind activities with the expression of one-dimensional attention shift probability vectors.Thus the randomness in attention shifts in actual thinking processes could be better expressed in the improved model.Besides,this paper also improved the calculation of attribute weights to make it more realistically significant.The improved equation could not only express the effects that fluctuations of attribute values have on the attribute attentions,but also show the thinking habit that passengers would like to pay attentions on the advantage attributes of the schemes.The dynamic choice experiments were implemented under the model frame with experimental psychology methods.Bi-alternative choice scenario experiment and multialternative choice scenario experiment are constructed for two groups of passengers under different trip purposes.The choice results and corresponding key parameters for modelling are acquired with survey data.In the parameter experiments,the tests for preference memorial effects are improved: separated pages are used in the tests to avoid the disturbances caused by the previous choice result,so that it could get closer to the implication of “memories”.Besides,by putting the test procedure at the end of the second section of time countdown,and by adding the verbal instructions of “immediately confirming” constraint,the test could better fulfill the rule of continuous in preference of decision field model.In the choice experiments,a two stage decision test is designed for the pre-travel stage and in-travel stage of multi-mode transit trip to mimic the dynamic process in stages of transit trips.Two decision sections with continuous time countdown are implemented in each stage to mimic the micro processes of passengers' preference evolutions in different thinking time sections.Meanwhile,the choice aversion phenomenon is captured in the evolution processes.Basing on the results of the experiments,four groups of Bi-alternative choice scenario simulations were implemented towards the different passenger groups under various travel scenarios.The micro psychological status that human have in thinking process is intuitively expressed from the perspective of attribute attention shifts,attribute valence distributions and preference evolution characteristics,and the mechanism of decision formulation is explained.According to the statistical analysis of the decision results,comparing to the basic model,the improved model that considers both characteristics of decision makers and heterogeneity of schemes shows better performance in describing the decision making results for the passengers in different groups and under different trip scenarios.By comparing the effective decision time and preference intensities between the two groups,it was found that weak memory effect group has slower speed in preference evolution and lower preference intensity in decision making than the strong memory effect group.Thus,the method for the setting of preference intensity threshold value is proposed: The preference intensity threshold value should be set according to the overall distributions of preferences that is shown in the effective decision time sections.The proposed method could avoid the mismatching of threshold value between the existing methodology and practical use.The results of simulations and experiments show that the choice reversals phenomenon is caused by the time-varying fluctuations in preference intensities in the process of thinking.The different preferences caused by evolution degree differences in different time sections could bring different choice results.The weak memory effects have higher frequencies in the fluctuations than the strong memory effects and are easier to occur choice reversals.By comparing the results of three-alternative simulations to the bi-alternative simulations,it shows that the two simulations have similar mechanisms of preference evolution.However,three kinds of phenomenon that violates the decision axioms were found in the threealternative simulations,namely attraction effects,compromise effects and similarity effects.The results of micro preference evolution analysis show that the heterogeneities of schemes have key impacts on these irrational phenomenon.The new schemes triggered the fluctuations in the initial attention on attributes and in the traffic statuses,and they leaded to the changes in attribute attention distributions,so that new trend in preference evolution were produced.A case study is carried out to study passengers' bus travel decision behavior before and after the opening of Chengdu Metro Line 3 Phase 3.Decision field model is adopted to analyze passengers' cross-regional travel decision process,intra-regional travel decision process,and the two-stage dynamic decision making process under fluctuated traffic conditions.The characteristics of passengers' decision-making behavior in real bus travel is obtained for residents along the line,and the corresponding optimization and adjustment measures were proposed to provide guidance for the coordination and integration of the operation of subway and bus.