Research On Data-driven Consistency Algorithm For Multi-agent Systems

Data-driven control is a control algorithm independent of the dynamics of the system and only a control method based on input and output data.As a kind of data-driven control algorithm,gradient-descent-based model-free adaptive control(GDBMFAC)method has broad application prospects in recent years.However,GDBMFAC often has the disadvantages of slow convergent rate and low tracking accuracy in practical consistency control system.In order to optimize the basic model-free adaptive control algorithm,the following research has been conducted:1.This paper collects and sorts out various improved gradient descent methods proposed in the current literature,and applies them to the field of model-free adaptive control.The feasibility and effectiveness of various optimization algorithms have been compared through simulation.2.In this paper,a model-free adaptive control algorithm based on adaptive momentum estimation method with weight decay(Adam WBMFAC)has been proposed.Compared with GDBMFAC,this algorithm can greatly accelerate the convergent rate and improve the tracking accuracy.3.This paper considers an actual multi-agent system with four permanent magnet linear motors Simulation results show that Adam WBMFAC has significant advantages over GDBMFAC.4.Aiming at the time delay,network attack and other phenomena that often exist in the actual network systems,the proposed algorithm is further improved to make it more universal.The research results show that the proposed algorithm is superior to other model-free adaptive control algorithms based on other improved gradient descent methods.