Distributed stochastic routing optimization using expander graph theory

In this paper, we propose a novel algorithm for distributed and decentralized stochastic routing to optimize the convergence rate and the resilience of routing. More precisely, the novelty of our method is the provision of a stochastic routing technique which greatly improves the spectral gap of the routing matrix, as compared to existing methods. We define a novel measure of effective expansion capability of a node, which we use to maximize the spectral gap of the stochastic routing matrix. Simulation results demonstrate that our proposed routing method significantly improves the convergence rate and resilience of routing compared to other existing methods.