Aiming at characteristics of non-linear temperature and large time delay of regenerative oxidation furnace

as well as the current situation that the traditional predictive control algorithm is difficult to establish an accurate mathematical model and the control precision is not high

a predictive control algorithm for non-linear system based on extra trees（ET） is proposed.The algorithm trains and studies the historical input and output data

establishes furnace temperature prediction model

and realizes the accurate prediction of furnace temperature.At the same time

rolling optimization controller is designed based on the algorithm of beetle antenna search（BAS）.The simulation results indicate that compared with the traditional predictive control method

the predictive control method based on BAS has a predictive precision of ±1.5 ℃ and a control precision of ±0.5 ℃.The proposed method has shorter rise time

smaller overshoot and smaller steady-state errors.