Broad Learning Aided Model Predictive Control With Application to Continuous Stirred Tank Heater

This paper presents a data-driven predictive controller based on the broad learning algorithm without any prior knowledge of system model. The is realized by regressing model using online process data and incremental algorithm. proposed control (MPC) approach requires less computational load compared to other neural network MPC approaches. More importantly, precision enhanced with reduced operating an appropriate approximation proved have no influence convergence Compared partial form dynamic linearization aided free (PFDL-MFC), performance illustrated through continuous stirred tank heater (CSTH) benchmark.