Final Activity Report Summary - SAPAD (Self-Aware Networks, Performance and Adaptivity)
This is the final report on our research on self-aware adaptive quality of service (QoS) driven network systems and on experiments with the cognitive packet network (CPN), which adaptively selected paths so as to offer best effort QoS to the end users based on user defined QoS.
CPN used neural network based reinforcement learning to make routing decisions separately at each node. The project pursued both an experimental agenda, based on system design, implementation and experimentation, and the study of analytical and theoretical aspects concerning QoS driven routing. A novel extension used a genetic algorithm to generate and maintain paths from previously discovered information. It matched their fitness with respect to the desired QoS goals. We also developed and tested a new QoS-based algorithm to defend networks against distributed denial of service attacks. Finally, we developed a theoretical model that allowed for the computation of the packet travel times in wireless sensor networks based on a Brownian motion model.
CPN used neural network based reinforcement learning to make routing decisions separately at each node. The project pursued both an experimental agenda, based on system design, implementation and experimentation, and the study of analytical and theoretical aspects concerning QoS driven routing. A novel extension used a genetic algorithm to generate and maintain paths from previously discovered information. It matched their fitness with respect to the desired QoS goals. We also developed and tested a new QoS-based algorithm to defend networks against distributed denial of service attacks. Finally, we developed a theoretical model that allowed for the computation of the packet travel times in wireless sensor networks based on a Brownian motion model.