TY - GEN

T1 - Distributed constrained optimization over noisy networks

AU - Srivastava, Kunal

AU - Nedić, Angelia

AU - Stipanović, Dušan M.

PY - 2010

Y1 - 2010

N2 - In this paper we deal with two problems which are of great interest in the field of distributed decision making and control. The first problem we tackle is the problem of achieving consensus on a vector of local decision variables in a network of computational agents when the decision variables of each node are constrained to lie in a subset of the Euclidean space. We assume that the constraint sets for the local variables are private information for each node. We provide a distributed algorithm for the case when there is communication noise present in the network. We show that we can achieve almost sure convergence under certain assumptions. The second problem we discuss is the problem of distributed constrained optimization when the constraint sets are distributed over the agents. Furthermore our model incorporates the presence of noisy communication links and the presence of stochastic errors in the evaluation of subgradients of the local objective function. We establish sufficient conditions and provide an analysis guaranteeing the convergence of the algorithm to the optimal set with probability one.

AB - In this paper we deal with two problems which are of great interest in the field of distributed decision making and control. The first problem we tackle is the problem of achieving consensus on a vector of local decision variables in a network of computational agents when the decision variables of each node are constrained to lie in a subset of the Euclidean space. We assume that the constraint sets for the local variables are private information for each node. We provide a distributed algorithm for the case when there is communication noise present in the network. We show that we can achieve almost sure convergence under certain assumptions. The second problem we discuss is the problem of distributed constrained optimization when the constraint sets are distributed over the agents. Furthermore our model incorporates the presence of noisy communication links and the presence of stochastic errors in the evaluation of subgradients of the local objective function. We establish sufficient conditions and provide an analysis guaranteeing the convergence of the algorithm to the optimal set with probability one.

UR - http://www.scopus.com/inward/record.url?scp=79953144206&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=79953144206&partnerID=8YFLogxK

U2 - 10.1109/CDC.2010.5717947

DO - 10.1109/CDC.2010.5717947

M3 - Conference contribution

AN - SCOPUS:79953144206

SN - 9781424477456

T3 - Proceedings of the IEEE Conference on Decision and Control

SP - 1945

EP - 1950

BT - 2010 49th IEEE Conference on Decision and Control, CDC 2010

T2 - 2010 49th IEEE Conference on Decision and Control, CDC 2010

Y2 - 15 December 2010 through 17 December 2010

ER -