Iterative learning control algorithms based on complex stochastic distribution systems

In this paper, a new generalized iterative learning algorithm is first proposed based on complex non-Gaussion stochastic control systems. Following designed neural networks are used to approximate the output PDF of the stochastic system in the repetitive processes or the batch processes, the tracking control to PDF is transformed into a parameter adaptive tuning problem in NN basis function. Under this framework, we study a new model free iterative learning control problem and propose a convex optimization algorithm based on a set of designed LMIs and L1 performance index. Such an algorithm has the advantage of the improvement of the closed-loop output PDF tracking performance and robustness. Simulation results are given to demonstrate the effectiveness of the proposed approach.