Distributed Linear Parameter Estimation: Asymptotically Efficient Adaptive Strategies

The paper considers the problem of distributed adaptive linear parameter estimation in multi-agent inference networks. Local sensing model information is only partially available at the agents and inter-agent communication is assumed to be unpredictable. The paper develops a generic mixed time-scale stochastic procedure consisting of simultaneous distributed learning and estimation, in which the agents adaptively assess their relative observation quality over time and fuse the innovations accordingly. Under rather weak assumptions on the statistical model and the inter-agent communication, it is shown that, by properly tuning the consensus potential with respect to the innovation potential, the asymptotic information rate loss incurred in the learning process may be made negligible. As such, it is shown that the agent estimates are asymptotically efficient, in that their asymptotic covariance coincides with that of a centralized estimator (the inverse of the centralized Fisher information rate for Gaussian systems) with perfect global model information and having access to all observations at all times. The proof techniques are mainly based on convergence arguments for non-Markovian mixed time scale stochastic approximation procedures. Several approximation results developed in the process are of independent interest.

[1]  V. Fabian On Asymptotic Normality in Stochastic Approximation , 1968 .

[2]  Gang George Yin,et al.  Consensus Formation in a Two-Time-Scale Markovian System , 2009, Multiscale Model. Simul..

[3]  Lihua Xie,et al.  Distributed Consensus With Limited Communication Data Rate , 2011, IEEE Transactions on Automatic Control.

[4]  Andrey V. Savkin,et al.  The problem of state estimation via asynchronous communication channels with irregular transmission times , 2003, IEEE Trans. Autom. Control..

[5]  Ioannis D. Schizas,et al.  Stability analysis of the consensus-based distributed LMS algorithm , 2008, 2008 IEEE International Conference on Acoustics, Speech and Signal Processing.

[6]  Srdjan S. Stankovic,et al.  Decentralized Parameter Estimation by Consensus Based Stochastic Approximation , 2007, IEEE Transactions on Automatic Control.

[7]  John N. Tsitsiklis,et al.  Convergence theories of distributed iterative processes: A survey , 1986 .

[8]  Angelia Nedic,et al.  Incremental Stochastic Subgradient Algorithms for Convex Optimization , 2008, SIAM J. Optim..

[9]  Reza Olfati-Saber,et al.  Kalman-Consensus Filter : Optimality, stability, and performance , 2009, Proceedings of the 48h IEEE Conference on Decision and Control (CDC) held jointly with 2009 28th Chinese Control Conference.

[10]  Ruggero Carli,et al.  Distributed Kalman filtering using consensus strategies , 2007, 2007 46th IEEE Conference on Decision and Control.

[11]  José M. F. Moura,et al.  Distributing the Kalman Filter for Large-Scale Systems , 2007, IEEE Transactions on Signal Processing.

[12]  Le Yi Wang,et al.  Asymptotic properties of consensus-type algorithms for networked systems with regime-switching topologies , 2011, Autom..

[13]  R.M. Murray,et al.  On a decentralized active sensing strategy using mobile sensor platforms in a network , 2004, 2004 43rd IEEE Conference on Decision and Control (CDC) (IEEE Cat. No.04CH37601).

[14]  Ali H. Sayed,et al.  Diffusion Least-Mean Squares Over Adaptive Networks: Formulation and Performance Analysis , 2008, IEEE Transactions on Signal Processing.

[15]  O. Nelles,et al.  An Introduction to Optimization , 1996, IEEE Antennas and Propagation Magazine.

[16]  José M. F. Moura,et al.  Cooperative Convex Optimization in Networked Systems: Augmented Lagrangian Algorithms With Directed Gossip Communication , 2010, IEEE Transactions on Signal Processing.

[17]  T. Lai,et al.  Asymptotically efficient self-tuning regulators , 1987 .

[18]  S. Mitter,et al.  Recursive stochastic algorithms for global optimization in R d , 1991 .

[19]  José M. F. Moura,et al.  Weight Optimization for Consensus Algorithms With Correlated Switching Topology , 2009, IEEE Transactions on Signal Processing.

[20]  Soummya Kar,et al.  Distributed Parameter Estimation in Sensor Networks: Nonlinear Observation Models and Imperfect Communication , 2008, IEEE Transactions on Information Theory.

[21]  Ali Jadbabaie,et al.  Non-Bayesian Social Learning , 2011, Games Econ. Behav..

[22]  Reza Olfati-Saber,et al.  Consensus and Cooperation in Networked Multi-Agent Systems , 2007, Proceedings of the IEEE.

[23]  G. Yin,et al.  Discrete-Time Markov Chains: Two-Time-Scale Methods and Applications , 2004 .

[24]  Asuman E. Ozdaglar,et al.  Distributed Subgradient Methods for Convex Optimization Over Random Networks , 2011, IEEE Transactions on Automatic Control.

[25]  John Baillieul,et al.  Robust and efficient quantization and coding for control of multidimensional linear systems under data rate constraints , 2006, CDC 2006.

[26]  Stephen P. Boyd,et al.  Randomized gossip algorithms , 2006, IEEE Transactions on Information Theory.

[27]  V. Borkar Stochastic Approximation: A Dynamical Systems Viewpoint , 2008 .

[28]  V. Fabian Stochastic Approximation of Minima with Improved Asymptotic Speed , 1967 .

[29]  Jonathan H. Manton,et al.  Stochastic approximation for consensus seeking: Mean square and almost sure convergence , 2007, 2007 46th IEEE Conference on Decision and Control.

[30]  John N. Tsitsiklis,et al.  Distributed Asynchronous Deterministic and Stochastic Gradient Optimization Algorithms , 1984, 1984 American Control Conference.

[31]  Alejandro Ribeiro,et al.  Consensus in Ad Hoc WSNs With Noisy Links—Part I: Distributed Estimation of Deterministic Signals , 2008, IEEE Transactions on Signal Processing.

[32]  Tze Leung Lai,et al.  Asymptotic Properties of Nonlinear Least Squares Estimates in Stochastic Regression Models , 1994 .

[33]  Soummya Kar,et al.  Distributed Consensus Algorithms in Sensor Networks With Imperfect Communication: Link Failures and Channel Noise , 2007, IEEE Transactions on Signal Processing.

[34]  David A. Freedman,et al.  A Sharper Form of the Borel-Cantelli Lemma and the Strong Law , 1965 .

[35]  U. Feige,et al.  Spectral Graph Theory , 2015 .

[36]  H. Kushner,et al.  Asymptotic properties of distributed and communication stochastic approximation algorithms , 1987 .

[37]  R. O. Saber Consensus and cooperation in networked multi-Agent systems , 2007 .

[38]  J. K. Hunter,et al.  Measure Theory , 2007 .

[39]  Mikhail Borisovich Nevelʹson,et al.  Stochastic Approximation and Recursive Estimation , 1976 .

[40]  Sekhar Tatikonda,et al.  Control under communication constraints , 2004, IEEE Transactions on Automatic Control.

[41]  Soummya Kar,et al.  Gossip Algorithms for Distributed Signal Processing , 2010, Proceedings of the IEEE.

[42]  Soummya Kar,et al.  Convergence Rate Analysis of Distributed Gossip (Linear Parameter) Estimation: Fundamental Limits and Tradeoffs , 2010, IEEE Journal of Selected Topics in Signal Processing.

[43]  R. Bass Convergence of probability measures , 2011 .

[44]  Stephen P. Boyd,et al.  A scheme for robust distributed sensor fusion based on average consensus , 2005, IPSN 2005. Fourth International Symposium on Information Processing in Sensor Networks, 2005..