Approximate consensus in the dynamic stochastic network with incomplete information and measurement delays

Consideration was given to the problem of achieving an approximate consensus in the decentralized stochastic dynamic network under incomplete information about the current states of the nodes, measurement delay, and variable structure of links. Solution was based on the protocol of local voting with nonvanishing steps. It was proposed to analyze dynamics of the closed network with the use of the method of averaged models which was extended to the systems with measurement delays. This method enables one to establish good analytical estimates of the permissible length of the step providing the desired accuracy of consensus and appreciably reduce the computational burden of simulation. The results obtained were applied to the analysis of the dynamics of the system of balancing the computer network loading.