Abstract:         

Routing of data through a relay network over long distances has been considered a major challenge over the years since the network can be compromised by random unknown attackers and inadvertent configurations. An important research challenge is to find the optimal placement of relay nodes in terms of power consumption, higher throughput, and security against malicious attackers. In a traditional communication network architecture, the relay nodes are static in nature. In our work we proposed a dynamic nature of devices that are used in communication. In this work, our objective is to establish an efficient distributed communication network of mobile devices that aims to maximize the network throughput while effectively resisting the attacks of malicious jammers. We consider a multi-agent framework in which the mobile devices act as relay nodes, whereas the base-stations are static. There are jammers in the field that actively try to cease and deteriorate communication by injecting noise. Thus, our objective is to find optimal locations of relay nodes at which the impact of the jammers is minimized. Moreover, at these optimal locations, the relay network should efficiently establish communication among the base stations. While computing these optimal locations, we will incorporate the energy cost associated with mobility as well as communication. Therefore, we will have a multi objective optimization problem for which we will explore techniques from Multi-Agent Reinforcement learning (MARL) and game theoretic methods.

Evaluation Committee:

• Supervisor:         Dr. Hassan Jaleel
• Evaluator:          Dr. Talha Manzoor