1st International Conference on Game Theory for Networks

Research Article

Asynchronous gossip algorithms for stochastic optimization

  • @INPROCEEDINGS{10.1109/GAMENETS.2009.5137386,
        author={S. Sundhar Ram and Angelia Nedi and Venugopal  Veeravalli},
        title={Asynchronous gossip algorithms for stochastic optimization},
        proceedings={1st International Conference on Game Theory for Networks},
        publisher={IEEE},
        proceedings_a={GAMENETS},
        year={2009},
        month={6},
        keywords={},
        doi={10.1109/GAMENETS.2009.5137386}
    }
    
  • S. Sundhar Ram
    Angelia Nedi
    Venugopal Veeravalli
    Year: 2009
    Asynchronous gossip algorithms for stochastic optimization
    GAMENETS
    IEEE
    DOI: 10.1109/GAMENETS.2009.5137386
S. Sundhar Ram1, Angelia Nedi1, Venugopal Veeravalli1
  • 1: Electr. & Comput. Eng. Dept., Univ. of Ilinois, Urbana, IL, USA

Abstract

We consider a distributed multi-agent network system where the goal is to minimize an objective function that can be written as the sum of component functions, each of which is known (with stochastic errors) to a specific network agent. We propose an asynchronous algorithm that is motivated by a random gossip scheme where each agent has a local Poisson clock. At each tick of its local clock, the agent averages its estimate with a randomly chosen neighbor and adjusts the average using the gradient of its local function that is computed with stochastic errors.We investigate the convergence properties of the algorithm for two different classes of functions: differentiable but not necessarily convex and convex but not necessarily differentiable.