Vehicle Platooning combines lower fuel consumption with safe and efficient transportation in a cooperative driving application. To ensure this, vehicles must communicate quickly and reliably. The sliding window protocol Caterpillar Random Linear Network Coding (CRLNC) is used as it does not rely on acknowledgements and uses network coding to reduce losses. However, the protocol was intentionally designed for single hop scenarios. Therefore, three contributions to the CRLNC protocol are presented in this paper to optimize it for multi hop/single path transmissions. The contributions are idle-slot-management, adaptive-window, and queue-management. By effectively combining these contributions, a large gain in throughput of up to 40% is achieved. Furthermore, it was investigated what are the optimal protocol parameters for CRLNC in single path scenarios are. It is shown that the best transmission in terms of throughput and delay is achieved, when the code rate is less than one minus the error probability and a large window size is used. Even if the error probability is unknown, a large window size will bring more benefits to the transmission.

Worldwide Interoperability for Microwave Access (WiMAX) mobile communication systems deliver all services through the Internet Protocol. Therefore, an important and crucial part of such systems is the Quality of Service (QoS) support to fulfill the needs of real-time applications. The underlying IEEE 802.16 standard specifies the QoS entities and signaling, but all algorithms are left open for vendor differentiation. In this paper, we propose a Mobile WiMAX QoS concept. It consists of a QoS aware traffic shaping and scheduling algorithm. In contrast to other proposals, our scheduling algorithm does not differentiate between service classes, but between the guaranteed and sustained traffic rate specified by the IEEE 802.16-2012 standard. Our Two-Rate-Based Scheduler (TRBS) additionally accomplishes fairness based on radio resources between users after fulfilling guaranteed demands. Large scale simulation results show that our TRBS algorithm satisfies the QoS requirements of real-time flows. The comparison to Proportional Fair scheduling by means of the number of supported VoIP users and system goodput reveals a significantly better system performance of our scheduling algorithm in highly loaded cells.

Traffic demand of backbone networks will increase rapidly in the next years, while equipment costs per bit are shrinking. To cope with this development the migration of those backbones towards Carrier Ethernet is investigated. In this paper we discuss a Particle Swarm Optimization (PSO) heuristic that can be used to enhance the overall migration process. We will introduce the basic migration problem, as well as system models for a possible migration case. Improved heuristics for basic PSO and their application to network migration will be presented. Results regarding algorithm performance and optimized migration cost will be delivered in the paper.

This paper presents a performance study of two new proposed strategies for the reorganization of a cluster-tree and its address space: proactive and reactive algorithms. The goal of the proactive reconfiguration is to minimize the average number of hops between the possible source and destination pairs. It is achieved by attempting to reconnect nodes to a parent node with the highest possible level. The goal of the reactive reconfiguration is to reorganize the cluster on- demand when a node wishing to join the cluster cannot do it, because the possible parents exhausted their free address space. Our simulation results show that with an optimized reconfiguration a higher performance in terms of connection time and connectivity of the cluster-tree can be achieved.

We consider the general problem of the simulation of highly reliable systems operating in the presence of Gaussian noise. Our methodology uses importance sampling which has been shown to be a particularly effective method in the general discipline of rare-event simulation. The methods we propose are optimal in a certain sense, i.e., they are efficient. We also give a new class of simulation distributions that are universally efficient in the sense that they depend only on a single scalar parameter, regardless of the dimensionality of the underlying system or of the error sets to be simulated.