||The developments in information technology and e-commerce have created new opportunities for logistics service operators. Electronic information exchanges cause different parties in a global distribution network to coordinate activities and share resources easily, while real-time tracking of shipment and equipment status allows the planning and control of the resources to be more responsive to changes. On the other hand, managing resources effectively is increasingly complex and difficult because of the uncertainty which lies in demands, costs or transportation times. This thesis is motivated to model, analyze, and provide a strategy on managing the flow dynamically by leveraging the capability of obtaining real-time information. In particular, we look at the flow management problems appearing in three geographic levels, namely the global level, the regional level and the local level. In the global level, we study the intermodal transportation coordination between suppliers and customers with a consolidation center. We prove that this problem is NP-complete and then develop a set of heuristics to obtain some good solutions efficiently. In the regional level, we focus on the flow management problem with uncertain demands, with application to the drayage operation problem and other types of resource allocation problems. We formulate the problem as a multistage stochastic network with random arc capacities to minimize the expected cost. By taking advantage of the problem structure, we develop a decomposition solution method which provides a new upper bound for the expected total cost. In the local level, we introduce a new problem class, which we refer to as the K-shortest dynamic path problem, to capture the dynamism of the routing decision process under uncertain environment while considering the needs of meeting other concerns and criteria in practice through providing multiple paths.