Transit Assignment Under Uncertain Demand

The transit passenger demand is affected by many uncertainty factors, such as the change of economy, population and land use. Thus the passenger demand from the origin stop to the destination stop is uncertain. In order to aquire accurate estimates of the transit flows, the uncertain demand must be fully taken into account.This dissertaion introduced research methods to the processing of uncertain passenger demand firstly. Then the demand is classified into long term and short term uncertain demand. The dissertation mainly studuies on the passengers'route choice behaviors under the short term demand uncertainty; the corresponding transit equilibrium assignment model is formulated. This work is of great significant to the transit operations management. Main contents of the dissertation include the following aspects:The models and algorithms of transit assignment and uncertain demand processing are reviewed. Firstly, three typical of classic transit equilibrium assignment models with constant demand are introduced. Then, analytical and simulation methods are put forward for the processing of uncertain demand. The applicable conditions of each method are presented.When the long term transit assignment model with demand uncertainty is studied, there is no need to consider passengers'special travel behavior. In this situation, an actual link cost of transit network is affected by the impact of congested passenger flows; the number of passenger flows is restricted by the limited capacies of buses. For the purpose of the comprehensive expression of the travel cost, a transit equilibrium assignment with congestion effects and capacity constraint is foumulated. A solution algorithm for the model is applied in a numerical example. This model can be a guide for the work of long term transit planning.Supposed that passengers have the historiccal travel experience in the rush hour of short term, and then they make the travel decision with different risk preference attitudes. A risk preference based transit equilibrium assignment model is formulated. In the model, transit passengers have a probability of arriving late at their destination. It is assumed that each passenger has budgeted the travel time of each transit route, an approach is described to penalise late arrival under fixed departure times, this late penalty is used to describe passengers'risk preference decision. Passengers' risk-averse and risk-prone attitudes are generated under different conditions. Then, the specific transit route cost is deduced. At last the corresponding transit assignment model is applied to the condition of short term uncertain demand. The comparison of the three methods for the uncertain demand processing is presented in the application.Finally the short term transit assignment model based on the risk preference is applied to the transit operarions management, spesificly to the determination of transit network reliability, which can be a performance evaluation index of transit operation and management.