Research On Traffic Evolution Based On Self-organizing And Edge Of Chaos Decisions

Traffic assignment is one of the most fundamental problems in transportation research field,route choice problems and travel mode choice optimization problems are based on the basic traffic assignment model.The traditional study laid more emphasis on the assignment of traffic among different routes(travel modes)and the equilibrium state of the model,in which the complexity of travelers' behavior and the depiction of travelers' heterogeneity is lacking.This dissertation is the inheritance and development of the traditional traffic assignment model.Basing on the evolution of complex system,the focus of this research is the seeking of the relationship between the individual behavior and overall performance of the transportation system.This dissertation can be divided into the following sections.In chapter 2,the traffic assignment problems,their sub problems and the related research of complex system are reviewed and summarized,what should be improved based on the existing research is given,on this basis,the research framework and technology roadmap is constructed.Furthermore,the features of multi-agent and self-organization in transportation system are analyzed,the framework of multi-agent model based on complex transportation system is constructed,and the effect of non-rational elements on travelers' group decision is stated.The shortcomings and merits of the traditional traffic assignment modeling idea are proposed through the analysis of different types of decision-making model,which illustrates the suitability and reasonability of introducing self-organizing multi-agent travelers into the modeling of complex transportation system.In chapter 3,the self-organizing multi-agent traffic evolution model is established by introducing a variety of travelers' attributes and behavioral rules based on cellular automata,the effect of travelers' social interaction(intensity and range)on the performance of the road network is analyzed.Moreover,based on the conception of"edge of chaos" in complexity science,the effect of travelers' behaviors on the transportation system is quantified and measured through the introduction of "Langton"parameter and the change of entropy.Optimization results of traffic assignment based on the state of "edge of chaos" is obtained.The new model inherits the characteristics of traditional model,as well as reveals the relationship between travelers' information level and traffic efficiency.In chapter 4,as a very common mental activity in our daily life,travelers' emotion and the idea of social emotion computing are introduced in travelers' self-organizing route choice behaviors.The evolution rule of "cellular genetic algorithm",which is closer to travelers' behavior,is used to depict travelers' path optimization,based on this,the new multi-agent traffic evolution model is established,the effect of travelers'behavior on the performance of the road network is analyzed.By designing travelers'"decision coefficient" and "evolution entropy of route choice behavior",the performance of the road network when the transportation system is in the "edge of chaos" state is calculated and travelers' behavioral complexity in this process is analyzed.In chapter 5,the cumulative prospect theory is introduced to model travelers'perceptions of utility.As the key parameter,travelers' reference point is studied and the heterogeneous reference points with the characteristic of dynamic evolution are modeled by establishing the small-world agent social network that is closer to real world.The reference point-dependent effects and the effect of travelers' heterogeneous level on route traffic are analyzed.By introducing the "decision coefficient" and"evolution entropy of route choice behavior",the road network's performance and the relationship between travelers' reference points and evolution complexity are calculated when the transportation system is in the state of "edge of chaos".In chapter 6,the multi-agent self-organizing traffic evolution and cumulative prospect theory are introduced into the modeling of travelers' choice among different travel modes,in which the formation of transportation demand with more detailed psychology pattern is designed,based on this,optimization of travel fare is studied,effect of travelers' decision attributes on travel cost and traffic flow is analyzed.By introducing the "decision coefficient" and "evolution entropy of mode choice behavior",optimization result of parameters and equilibrium price when the travelers are in the state of "edge of chaos" are calculated,the relationship between travelers' reference points and evolutionary complexity are analyzed.In this dissertation,travel behavior is modeled based on the reasonable assumptions from the perspective of complex system and the science of complexity,which effectively enriched the research idea and method of traditional traffic assignment model.Furthermore,the research of this dissertation can help operators giving decision support to individual travelers to make the transportation system more efficient.