Model-Free Optimal Output Cluster Synchronization Control for Multiagent Systems
Abstract: In this article, the model-free optimal output cluster
synchronization control problem is investigated for nonlinear
multiagent systems (MASs). First, in view of the unknown output
of leader, relying on practical prescribed-time performance
function, an observer is designed for each follower to estimate
the output of leader, and can achieve the desired accuracy
within prescribed time. Then, based on the designed observer, an
augmented system consisting of observer dynamics and follower
dynamics is constructed and the cost functin is built for each
follower. Accordingly, the optimal output cluster synchronization
control problem is transformed into a numerical solution to solve
the Hamilton-Jacobian-Bellman equation (HJBE). Subsequently,
the off-policy reinforcement learning (RL) algorithm is addressed
to learn the solution to HJBE without any knowledge of the
system dynamics. Meanwhile, to release computational burden,
the single critic neural network (NN) framework is employed to
implement the algorithm, where the least square method is used
for training the NN weights. Thus, the designed control algorithm
can minimize the cost functions and ensure the output cluster
synchronization of MASs with the unknown system dynamics
and unavailable leader output. Finally, the simulation examples
confirm the validity of the designed control scheme.
Submission Number: 240
Loading