Asynchronous latency analysis on decentralized iterative algorithms for large scale networked systems

Decentralized methods are often desirable for solving many control and optimization problems for large scale networked systems. Convergence of the decentralized iterative method may be affected by the delay caused by communication and computation. This paper investigates the latency impact of asynchronous decentralized algorithms with the contractive property on convergence. An asynchronous decentralized algorithm is presented with its convergence conditions and the theoretical upper bound of latency is derived. Numerical examples are shown to demonstrate the effectiveness of the algorithm, and the testing results are analyzed.