Randomized consensus algorithms over large scale networks

被引:215
作者
Fagnani, Fabio [1 ]
Zampieri, Sandro [2 ]
机构
[1] Politecn Torino, Dipartmento Matemat, I-10129 Turin, Italy
[2] Univ Padua, Dept Informat Engn, I-35131 Padua, Italy
关键词
consensus algorithms; large-scale networks; random algorithms; concentration of performance; mean square analysis;
D O I
10.1109/JSAC.2008.080506
中图分类号
TM [电工技术]; TN [电子技术、通信技术];
学科分类号
0808 ; 0809 ;
摘要
Various randomized consensus algorithms have been proposed in the literature. In some case randomness is due to the choice of a randomized network communication protocol. In other cases, randomness is simply caused by the potential unpredictability of the environment in which the distributed consensus algorithm is implemented. Conditions ensuring the convergence of these algorithms have already been proposed in the literature. As far as the rate of convergence of such algorithms, two approaches can be proposed. One is based on a mean square analysis, while a second is based on the concept of Lyapunov exponent. In this paper, by some concentration results, we prove that the mean square convergence analysis is the right approach when the number of agents is large. Differently from the existing literature, in this paper we do not stick to average preserving algorithms. Instead, we allow to reach consensus at a point which may differ from the average of the initial states. The advantage of such algorithms is that they do not require bidirectional communication among agents and thus they apply to more general contexts. Moreover, in many important contexts it is possible to prove that the displacement from the initial average tends to zero, when the number of agents goes to infinity.
引用
收藏
页码:634 / 649
页数:16
相关论文
共 28 条
[1]  
Alon N., 2004, The probabilistic method
[2]  
[Anonymous], 1979, ERGODIC THEORY DIFFE
[3]  
Arnold L., 1998, Springer Monographs in Mathematics
[4]  
BLONDEL V, 2005, P JOINT IEEE CDC ECC, P2993
[5]   Randomized gossip algorithms [J].
Boyd, Stephen ;
Ghosh, Arpita ;
Prabhakar, Balaji ;
Shah, Devavrat .
IEEE TRANSACTIONS ON INFORMATION THEORY, 2006, 52 (06) :2508-2530
[6]   Communication constraints in the average consensus problem [J].
Carli, Ruggero ;
Fagnani, Fabio ;
Speranzon, Alberto ;
Zampieri, Sandro .
AUTOMATICA, 2008, 44 (03) :671-684
[7]  
COGBURN R, 1986, CONT MATH, V50, P199
[8]   DYNAMIC LOAD BALANCING FOR DISTRIBUTED MEMORY MULTIPROCESSORS [J].
CYBENKO, G .
JOURNAL OF PARALLEL AND DISTRIBUTED COMPUTING, 1989, 7 (02) :279-301
[9]   Efficient schemes for nearest neighbor load balancing [J].
Diekmann, R ;
Frommer, A ;
Monien, B .
PARALLEL COMPUTING, 1999, 25 (07) :789-812
[10]  
DYMAKIS ASA, 2006, GEOGRAPHIC GOSSIP EF