The formal definition of network coding by Ahlswede et al. in 2000 has led to several breakthroughs in information theory, for example solving the bottleneck problem in butterfly networks and breaking the min-cut max-flow theorem for multicast communication. Especially promising is the usage of network coding as a packet-level Forward Error Correction (FEC) scheme to increase the robustness of a data stream against packet loss, also known as intra-session coding. Yet, despite these benefits, network coding-based FEC is still rarely deployed in real-world networks. To bridge this gap between information theory and real-world usage, this cumulative thesis will present our contributions to the integration, evaluation, and optimization of network coding-based FEC. The first set of contributions introduces and evaluates efficient ways to integrate coding into UDP-based IoT protocols to speed up bulk data transfers in lossy scenarios. This includes a packet-level FEC extension for the Constrained Application Protocol (CoAP) [P1] and one for MQTT for Sensor Networks (MQTT-SN), which levels the underlying publish-subscribe architecture [P2]. The second set of contributions addresses the development of novel evaluation tools and methods to better quantify possible coding gains. This includes link ’em, our award-winning link emulation bridge for reproducible networking research [P3], and also SPQER, a word recognition-based metric to evaluate the impact of packet loss on the Quality of Experience of Voice over IP applications [P5]. Finally, we highlight the impact of padding overhead for applications with heterogeneous packet lengths [P6] and introduce a novel packet-preserving coding scheme to significantly reduce this problem [P4]. Because many of the shown contributions can be applied to other areas of network coding research as well, this thesis does not only make meaningful contributions to specific network coding challenges, but also paves the way for future work to further close the gap between information theory and real-world usage.

Integrated production and distribution scheduling has emerged as an important research topic in the areas of combinatorial optimization and supply chain management. In this thesis we consider problems that occur frequently in the manufacturing of products with a short lifespan or in make-to-order business models, where the production is triggered only if a corresponding customer order is received. The focus is put on the coordination of production and distribution processes to meet the requirements of the customers as closely as possible while at the same time maximizing the revenues of the manufacturer. We analyze a basic production and distribution model and study several extensions in order to incorporate more restrictions from practice. These model extensions lead to complex scenarios that make the class of considered problems very challenging to solve. We address their complexity by developing appropriate heuristic solution methods based on meta-heuristics such as tabu search or iterated local search. Moreover, we evaluate the performance of the proposed algorithms on a set of test instances. Practical applications in which the previously mentioned problems need to be solved include the production and transportation of certain intermediate products like adhesive materials in chemical industry and the delivery of ready-mix concrete from a production plant to several construction sites by a fleet of delivery trucks.