Heuristically Speeding up Anonymous Dynamic Network Computations

Austin Halper, Dariusz R. Kowalski, Miguel A. Mosteiro

Abstract: Motivated by the current large gap between lower and upper bounds for Counting, i.e. the problem of computing the number of nodes, in Anonymous Dynamic Networks, we study heuristic modifications of the deterministic Methodical Multi-counting (MMC) protocol aimed to reduce its running time in practical applications. The heuristic presented here is the result of an extensive experimental study tuning all the parameters of the algorithm while maintaining correctness. We evaluate our heuristic MMC with thorough simulations on worst and average case topologies under different assumptions of dynamicity. Our results achieve speed-ups up to one or two orders of magnitude, which demonstrates that, as expected, the theoretical analysis overestimates the necessary running time to cope with adversarial scenarios, but in practice it is possible to tune the algorithm for specific topologies or other network characteristics without sacrificing correctness.