| Emerging multi-hop wireless networks provide a low-cost and flexible infrastructure that can be simultaneously utilized by multiple users for a variety of applications, including delay-sensitive applications, such as multimedia streaming, mission-critical applications, etc. However, this wireless infrastructure is often unreliable and provides dynamically varying resources with only limited QoS support.;To improve the performance of the delay-sensitive applications and to support timely reaction to the network dynamics, the multi-hop network needs to be composed of autonomic nodes (agents), which can adapt, make their own transmission decisions and negotiate their wireless resources based on their available local information. Current wireless networking research has focused on coping with the environment disturbances, such as variations (uncertainties) of the wireless channel (e.g. fading) or source (e.g. multimedia traffic) characteristics, while neglecting the coupling dynamics among nodes, due to the shared nature of the wireless spectrum. However, characterizing and learning the neighboring nodes' actions and the evolution of these actions over time is vital in order to construct an efficient and robust solution for delay-sensitive applications. Hence, we propose and analyze various interactive learning schemes for these agents to learn the network dynamics and, based on this knowledge, foresightedly adapt their cross-layer transmission decisions such that they can efficiently utilize the shared, time-varying network resources. We show that the foresighted decision making significantly improves the agents' utilities under a variety of dynamic network scenarios (e.g. multimedia streaming over WLAN, energy-efficient transmission in mobile ad hoc networks, joint route/channel selection in multi-hop cognitive radio networks) and various network topologies as compared to existing state-of-the-art solutions.;In conclusion, our research adds a new, "cognitive", dimension to existing multi-hop wireless networks that enables the autonomic nodes to dynamically forecast the expected response to network dynamics of neighboring nodes and evaluate how specific forms of explicit and implicit signaling impact the performance of delay-sensitive applications. |
