Cyclic Stochastic Approximation with Disturbance on Input in the Parameter Tracking Problem Based on a Multiagent Algorithm

We consider the possibilities of using cyclic stochastic approximation for solving optimization problems of a nonstationary functional of the data produced by distributed observers (sensors) under constraints on the possibilities of simultaneous communication between the observers themselves. We achieve tracking the optimal value of parameters up to a certain level of quality with a multi-agent algorithm. The efficiency of the proposed approach is illustrated by an example of modeling the process of tracking the trajectories of a group of moving objects using a set of spatially distributed sensors.