Towards a Theory of Self-Organizing Networks for Countering WMD Attacks

Task 1: Neighbor Discovery and Network Formation PI Ying will be leading on developing distributed beacon schemes and characterize fundamental performance bounds on beacon cycle, which is related to the number of colors needed to guarantee that the minimum distance between a pair of nodes using the same color is lower bounded by a required distance. This problem will be tackled using spatial point processes, random graphs for modeling the network and graph theory for designing the coloring scheme and understanding the fundamental performance bound. Task 2: Real-Time Communication under Severe Resource Constraints To mitigate the impacts of WMD attacks, besides beacon exchange, survivors need to transmit text, pictures, voices in real-time manner to report their status and information about damages and essential resources (e.g., available food, water, power supplies). Therefore, the network needs to support real-time D2D communications in a highly dynamic network with limited communication resources such as power. Remarkable progress has been made over the past several years on real time communication in wireless networks. However, most results are developed based on the premise that network topology is static, and stationary and ergodic traffic flows are injected persistently into the network. This premise clearly does not hold in our network, where the network topology constantly changes and people will be exchanging real-time information with small sizes. PI Ying will be leading the efforts on distributed algorithms that can support real-time D2D communications and work well in networks with topological dynamics and severe resource constraints such as power, bandwidth and limited network knowledge. In particular, PI Ying will lead on research on characterizing the deadline-constrained network throughput region using ideas from information theory and stochastic networks, deriving stochastic control algorithms that achieve the deadline-constrained throughput region in networks with traffic, channel and topology dynamics, and developing distributed implementation that is practically implementable will be developed.