切换网络下加速分布式在线加权对偶平均算法

ACCELERATE DISTRIBUTED ONLINE WEIGHTED DUAL AVERAGE ALGORITHM IN SWITCHED NETWORKS

研究了切换网络下加速分布式在线加权对偶平均算法,提出了A-DOWDA算法。首先利用加权因子对对偶变量进行加权,其次在有向切换网络是周期强连通,且对应的邻接矩阵是随机的而非双随机的条件下,加速了算法的收敛速率,最后通过数值实验验证了算法的可行性。

We studies the distributed online weighted dual average algorithm is accelerated under switched network and an A-DWDA algorithm is proposed. Firstly, weighting factors are used to weight dual variables. Secondly, the directed switched network is periodically strongly connected, and the corresponding adjacency matrix is stochastic rather than doubly stochastic, the convergence speed of the algorithm is accelerated. Finally, a numerical experiment is performed to verify the effectiveness of the proposed algorithm.