In this thesis we focus on two specific transportation systems, namely postal automation and baggage handling. Postal automation: During the last decades the volume of magazines, catalogs, and other plastic wrapped mail items that have to be processed by post sorting centers has increased considerably. In order to be able to handle the large volumes of mail, state-of-the-art post sorting centers are equipped with dedicated mail sorting machines. The throughput of a post sorting machine is defined as the number of sorted mail items divided by the time needed to sort them. In this thesis we consider large letters of A4 size envelopes. Such mail items are called flats. Briefly, a state-of-the-art automated flats sorting machine operates as follows: the flats are inserted into transport boxes by feeding devices; the boxes carry the pieces with constant speed and sort them into static destination bins according to the selected sorting scheme. The throughput of a basic system sketched above can be augmented by designing a system where the bottom part consisting of destination bins can move bidirectional with piecewise constant speed. The model of the flats sorting system is event-based, determined using simulation. In order to compute the speed of the bottom system that maximizes the throughput of this machine, we implement and compare different variants of optimal control with gradually decreasing complexity namely: optimal control with a piecewise constant speed on time intervals of variable length, optimal control with a piecewise constant speed on time intervals of constant length, optimal control with a constant speed, and model-based predictive control with a piecewise constant speed on time intervals of constant length. The considered control methods are compared for several scenarios. In this thesis we also discuss the influence of the structural changes on the throughput. In particular, we consider structural changes such as increasing the number of feeding devices and changing their inserting point around the transport boxes, and increasing the velocity of the transport boxes. Baggage handling: The continuous need for reduction of costs in the air transport industry and the rise of low-cost carriers require a cost effective operation of the airports. Modern baggage handling systems transport luggage in an automated way using destination coded vehicles (DCVs). These vehicles transport the bags at high speeds on a railway of tracks. The DCV-based baggage handling operates as follows: given a dynamic demand of bags and a buffer of empty DCVs for each loading station, together with the network of single-direction tracks, the route of each DCV has to be computed subject to operational and safety constraints such that each of the bags to be handled arrives at its given end point within a specific time window. Currently, the networks have simple structure, the DCVs being routed through the system using routing schemes based on preferred routes. These routing schemes can be adapted to respond on the occurrence of predefined events. However, the load patterns of the system are highly variable, depending on e.g. the season, time of the day, type of aircraft at each gate, or the number of passengers for each flight. Therefore, we do not consider predefined preferred routes, but instead we develop and compare efficient control methods to determine the optimal routing in case of dynamic demand. In particular, we consider predictive and heuristic approaches implemented in a centralized, decentralized, and distributed manner. Furthermore, in order to efficiently determine the route choice of each DCV we also propose a hierarchical control framework that consists of a two-level control structure with local switch controllers at the lowest level and one higher supervisory controller. In this control framework, switch controllers provide position instructions for each switch in the network. The collection of switch controllers is then supervised by a so-called network controller that mainly takes care of the flow instructions for the switch controllers. Computing the optimal route choice yields a nonlinear, nonconvex, mixed integer optimization problem. The computational efforts required to determine the optimal route choice are high, and therefore, solving this optimization problem becomes intractable in practice. Consequently, we also present an alternative approach for reducing the complexity of the computations by writing the nonlinear optimization problem as a mixed integer linear programming (MILP) problem. The advantage is that for MILP optimization problems solvers are available that allow us to efficiently compute the global optimal solution. The solution of the MILP problem can then be used directly or as an initial starting point for the original optimization problem. To assess the performance of the proposed control approaches and control frameworks, we consider a benchmark case study, in which the methods are compared for several scenarios.