Based on the Moving Horizon Estimation of the Nonlinear Parameters Optimization

In order to solve error between the system output values and the model estimation values in non-linear system. A new method, the particle swarm optimization & sequential quadratic programming (PSO-SQP), is proposed to realize the on-line optimization to nonlinear complex system. Firstly, the PSO-SQP algorithm is proposed to solve the slow search speed of the PSO and easy convergence to the local minimum points of SQP. Then the method that the unknown parameters and states are estimated by the moving window based on moving horizon estimation (MHE) is also described in the paper. Finally, the two examples are simulated to test those methods. It is validated that this method can cut down iteration times about 40%~60%, the optimization of this algorithm can be about up to 30% on average.