Stochastic Mirror Descent for Convex Optimization with Consensus Constraints

The mirror descent algorithm is known to be effective in applications where it is beneficial to adapt the mirror map to the underlying geometry of the optimization model.

However, the effect of mirror maps on the geometry of distributed optimization problems has not been previously addressed. 

In this talk we will study an exact distributed mirror descent algorithms in continuous-time under additive noise and present the settings that enable linear convergence rates.

Our analysis draws motivation from the augmented Lagrangian and its relation to gradient tracking. 

To further explore the benefits of mirror maps in a distributed setting we present a preconditioned variant of our algorithm with an additional mirror map over the Lagrangian dual variables. This allows our method to adapt to the geometry of the consensus manifold and leads to faster convergence. 

We illustrate the performance of the algorithms in convex settings both with and without constraints.

This is joint work with: Anastasia Borovykh, Nikolas Kantas,  and  Grigorios A. Pavliotis