Title:Noisy source location on a line

Abstract:We study the problem of locating the source of an epidemic diffusion process from a sparse set of sensors, under noise. In a graph $G=(V,E)$, an unknown source node $v^* \in V$ is drawn uniformly at random, and unknown edge weights $w(e)$ for $e\in E$, representing the propagation delays along the edges, are drawn independently from a Gaussian distribution of mean $1$ and variance $\sigma^2$. An algorithm then attempts to locate $v^*$ by picking sensor (also called query) nodes $s \in V$ and being told the length of the shortest path between $s$ and $v^*$ in graph $G$ weighted by $w$. We consider two settings: static, in which all query nodes must be decided in advance, and sequential, in which each query can depend on the results of the previous ones.
We characterize the query complexity when $G$ is an $n$-node path. In the static setting, $\Theta(n\sigma^2)$ queries are needed for $\sigma^2 \leq 1$, and $\Theta(n)$ for $\sigma^2 \geq 1$. In the sequential setting, somewhat surprisingly, only $\Theta(\log\log_{1/\sigma}n)$ are needed when $\sigma^2 \leq 1/2$, and $\Theta(\log \log n)+O_\sigma(1)$ when $\sigma^2 \geq 1/2$. This is the first mathematical study of source location under non-trivial amounts of noise.