From Consensus to Robust Randomized Algorithms: A Symmetrization Approach

This paper interprets and generalizes consensus-type algorithms as switching dynamics leading to symmetrization with respect to the actions of a finite group. Explicit convergence results are provided in a grouptheoretic formulation, both for deterministic and for stochastic dynamics. We show how the symmetrization framework directly extends the scope of consensustype algorithms and results to applications as diverse as consensus on probability distributions (either classical or quantum), computation of the discrete Fourier transform, uniform random state generation, and openloop disturbance rejection by quantum dynamical decoupling. This indicates a way to extend the desirable robustness of consensus-inspired algorithms to even more fields of application.