Abstract:
In this paper we consider distributed convex optimization over time-varying undirected graphs. We propose a linearized version of primarily averaged network dual ascent (...Show MoreMetadata
Abstract:
In this paper we consider distributed convex optimization over time-varying undirected graphs. We propose a linearized version of primarily averaged network dual ascent (PANDA) that keeps the advantages of PANDA while requiring less computational costs. The proposed method, economic primarily averaged network dual ascent (Eco-PANDA), provably converges at R-linear rate to the optimal point given that the agents' objective functions are strongly convex and have Lipschitz continuous gradients. Therefore, the method is competitive, in terms of type of rate, with both DIGing and PANDA. The proposed method halves the communication costs of methods like DIGing while still converging R-linearly and having the same per iterate complexity.
Published in: ICASSP 2019 - 2019 IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP)
Date of Conference: 12-17 May 2019
Date Added to IEEE Xplore: 17 April 2019
ISBN Information: