| In the traffic planning and management,the transportation network design is an indispensable and important link,since it has an important influence on the other stages of the traffic planning and management,and its rationality and reliability are the precondition to ensure the normal operation of the whole traffic system.More importantly,there is a complex relationship which promotes and restricts each other between the road network and other subsystems of the city.Traditional traffic planning theories and network design methods focus on the measurement of traveler mobility,such as moving speed and travel time.However,mobility-driven methods can induce the increase of travel demand and further lead to the generation of traffic congestion.In this context,the goal of minimizing the travel time of the system is not suitable for the accessibility-based planning process of transportation network,nor can it consider the problem of social equity.As we know,one of the significant aims of transportation network design,orgnization and management is to improve the service level of the transportation network and the accessibility of individual trips in a certain period.Therefore,the method for maximizing network accessibility expands the potential strategy scope of the traffic planning and transportation network design problem to some extent.Based on the idea of system optimization,taking accessibility measures as comprehensive evaluation indices of traffic network,and taking transportation network design methods as the main theme,in this thesis the theoretical researches on discrete transportation network design problems are successively carried out under demand certainty,under demand uncertainty or with the consideration of spatiotemporal attributions of activities.The achievements of above researches are applied in the optimization of road traffic management and rail transit organization.Specifically,the research includes the following contents.(1)Flow accessibility measure and space-time accessibility measure.The continuous increase of traffic flow will inevitably lead to contradictions between the travellers' expected travel time and traffic congestion.In this context,constrained by a flow-dependent travel time budget,this thesis first proposes the flow accessibility measure on a physical network,which reflects the balance between actual traffic flow and potential traffic demand.In view of the spatiotemporal attributes of travel activities,the space-time accessibility measure is based on the space-time network and constrained by the travel time budget related to the origin-destination pair and departure time,which reflects real-time accessibility of origin-destination pair on the transportation network.Although these two accessibility indicators have different emphasis and use environment,both can reflect the contradictory unity between the subjective accessibility and objective accessibility,and measure the coupling relationship between transportation demand and supply under certain social development conditions.(2)Discrete transportation network design under the deterministic or undeterministic demand.Under the deterministic demand,the discrete transportation network design problem is formulated as a nonlinear mixed-integer bi-level programming model to maximize the flow accessibility measure,in which the upper level determines network design strategies,and in response the lower level is the user equilibrium model under the travel time budget.Under the undeterministic demand,the robust discrete transportation network design problem is formulated as a two-stage stochastic programming model to maximize the expected flow accessibility measure,in which the first stage generates the network design strategies,and the second stage deals with the undeterministic demand by setting different scenarios and uses the user equilibrium model under the travel time budget to assign the realized demand.In order to solve these two models,a heuristic algorithm based on probabilistic search and implicit enumeration method is designed through analyzing the feature of the problem.Taking the Sioux Falls network and the San Diego highway network as examples,some numerical experiments are designed,and some sensitivity analyses of parameters are conducted.The experimental results show that the proposed methods are suitable for traffic networks with different scales,and the obtained network design scheme can significantly improve the accessibility level over the entire network.(3)Robust discrete transportation network design with the consideration of spatiotemporal attributions of activities.In view of the impact of time-dependent and scenario-dependent activities on the discrete transportation network design,the robust discrete transportation network design problem is formulated as a 0-1 linear programming model to maximize the expected space-time accessibility measure.By introducing Lagrangian multipliers,the coupling constraints between physical links and space-time arcs in the model can be relaxed,so that the model can be effectively decomposed.The decomposed submodels are the constrained shortest path problem and the 0-1 knapsack problem,respectively.For solving the difficulty caused by the high-dimensional decision variables,an efficient heuristic algorithm is designed.Taking Sioux Falls network and San Diego highway network as examples,some numerical experiments are designed,and the influences of time budgets,capital budgets and scenario probabilities on the network design scheme are discussed.The experimental results show that the proposed methods are suitable for traffic networks with different scales,and the obtained robust network design scheme can significantly improve the accessibility level over the entire network.(4)Reversible lane optimal setting on traffic networks.As an application of the discrete transportation network design methods,the reversible lane optimal setting problem is formulated as a nonlinear mixed-integer bi-level programming model to maximize the flow accessibility measure,in which the upper level determines reversible lane setting strategies,and in response the lower level is the user equilibrium model under the travel time budget.Compared with the discrete transportation network design model based on flow accessibility measure,the value range of decision variables in this model is expanded,which increases the difficulties to solve.In order to obtain the approximate solution of the model,a progressive local search algorithm is designed,namely,the integer constraints are relaxed first,and then the local search algorithm is used to search the integer solution.Taking Sioux Falls network and Beijing main road network as examples,some numerical experiments are designed.The experimental results show that the proposed methods are suitable for traffic networks with different scales,and the obtained reversible lane optimal setting scheme can significantly improve the accessibility level over the entire network.(5)Collaborative optimization of the last train timetable on the metro network.Studying the connection problem of last trains on the metro network from the perspective of space-time network is equivalent to studying the space-time network design problem.As an application of the discrete transportation network design methods,by creatively using network design ideas,the collaborative optimization problem of the last train timetable on the metro network is formulated as three 0-1 linear programming models to maximize different accessibility measures.By introducing Lagrangian multipliers,the coupling constraints between travelers and trains in the models can be relaxed,so that the models can be effectively decomposed.The decomposed submodels are respectively the shortest space-time path problem and the shortest path problem in the network with negative weights and without circles.Based on the decomposition,an efficient heuristic algorithm is designed.Taking a small-scale network and Beijing metro network as examples,some numerical experiments are designed.The experimental results show that the proposed methods are suitable for metro networks with different scales,and the obtained collaborative timetable scheme of last trains can significantly improve the accessibility level over the entire metro network during the last train period.This thesis includes 49 figures and 21 tables,and refers to 181 literatures. |
PDF Full Text Request