Research On Modeling The Activity-travel Assignment Problem With Bounded Rationality

The space-temporal distribution of traffic flow in urban transport network is resulted in travelers'choice on activity location, travel route, travel mode, departure time etc. with various attributes. This study aims to study the dynamic activity-travel assignment problem with the consideration of travelers' bounded rationality choice behavior and heterogeneity. Based on decision making theories, the micro-decision behavior is discussed; based on traffic assignment models, the macro flow distribution is studied; the heterogeneity of traverls is studied from various perspectives. The main contributions are concluded as follow:(1) A comprehensive description on the content, process and bouded rationality of activity-travel decision is made, and the influence factors on the decision process is analysed based on the attribution theory; as a result of the analysis, travelers have bounded rational behavior on the knowledge of decision sets, valuing alternatives, subjective perception on risk and non-best decision rule; different decision theories discuss one or some facets bounded rationality; and the various of objective alternatives and the difference inter individuals or intra individuals are the two main factors that have influence on the decision results.(2) As the complex attributes that an activity-travel pattern choice includes, to represent the full choice facets of a pattern, some supernetworks are applied, such as spatial-time expanded network, multi-modal spatial-time expanded network, activity-travel state augmented multi-modal supernetwork and multi-modal multi-state supernetwork; based on these supnetworks, the activity-travel assignment models are introduced by the classificiation from time horizon, specificly, if the study horizon is short than one day, such as morning peak hour, the activity-travel assignment model is proposed to analysis short-term traffic control problem; if the study horizon is one day, the activity-travel assignment model is proposed for long-term traffic planning.(3) As the analysis above, firstly, the activity-travel assignment problem under uncertainty is studied. Considering the bouded rationality from the utility and probability that travelers perceived, based on a simplified multi-state supemetowrk, in which one path through the network can be represented as a daily activity-travel pattern with space and temporal information, a static prospect-based user equilibrium model with endogenous reference is proposed to study travelers' activity-travel choice behavior when travel time and activity duration follow certain random distribution.(4) Secondly, the quasi-dynamic activity-travel assignment problem with the considertation of the constraints of traveler's perception is studied. Based on the reference-dependent theory, travelers' bounded rationality on reference-dependent judgement on alternatives' value is discussed. Assuming travelers are divided into various classes by different references, a quasi-dynamic stochastic user equilibrium model with endogeneous references is proposed with the assumption that travelers can finish their activity, route, departure time choices during one interval. The convergence and performance of the proposed model is analyzed by a numerial example.(5) Thirdly, allowing for the situation that travelers possess accurate information of the urban-transportation system due to modern technologies, the dynamic activity-travel assignment problem is studied. In view of traveler heterogeneity, individuals can be divided into multiple classes according to their reference points. A reference-dependent multi-class dynamic user equilibrium model for activity-travel scheduling is proposed. Moreover, comparative analyses are conducted on the equilibrium states between utility-maximization (no reference) and reference-dependency of exogenous and endogenous references. A numerical example regarding combined departure-time and mode choice for commuting is conducted to illustrate the proposed model. The simulated results indicate that reference points and loss aversion attitudes have significant effects on the choice of departure time and mode.(6) Finally, given there are majority activity-travel alternatives in real transport network, different travelers have difference choice sets. The dynamic activity-travel assignment problem with heterogeneity choice sets is studies. Based on bouded decision rule, a dynamic activity-travel assignment model with the consideration of heterogeneity choice sets and bouded rationality is proposed. Moreover, as a promising way to ease the aforementioned urban issues, carsharing has recently attracted increasing attention in both practice and academia. As a new mode choice as free-floating carsharing (FFC) is, it's obvious that travelers can be dividied into two classes by whether they have a memebership of carsharing program. A dynamic user equilibrium model of activity-travel scheduling that adds free-floating shared car as a transport mode alternative is also proposed. By extending the multi-state supernetwork representation, the use of shared cars can be consistently embedded into daily activity-travel patterns. The dynamic traffic flow and supply-demand interaction of shared cars can be formulated endogenously. A simulation-based route swapping method of path flow adjustment is adopted to solve the dynamic user equilibrium model. Numerical examples under various scenarios demonstrate that fleet size, distribution and pricing of rental and parking of shared car significantly influence both the choice of shared cars and activity-travel patterns.