ANFIS-Based Adaptive Model Predictive Control for Complex Nonlinear Dynamic Systems under Uncertainty

Complex nonlinear dynamic systems encountered in industrial automation, chemical processes, energy systems, and advanced manufacturing environments are frequently affected by model uncertainties, nonlinear behavior, parameter variations, and external disturbances. These characteristics often degrade the performance of conventional control methods and limit their ability to maintain stability and control accuracy under changing operating conditions. Consequently, the development of intelligent control strategies capable of adaptation, learning, and prediction has become an important challenge in modern control engineering. This study proposes an ANFIS-based Adaptive Model Predictive Control (AMPC) framework for the intelligent control of complex nonlinear dynamic systems operating under uncertain conditions. The proposed methodology integrates an Adaptive Neuro-Fuzzy Inference System (ANFIS) with Model Predictive Control (MPC) to combine nonlinear system learning capabilities with predictive optimization. ANFIS is employed to identify and approximate the nonlinear dynamics of the controlled process using operational data, while MPC predicts future system behavior and determines optimal control actions over a specified prediction horizon. This integrated structure enables continuous adaptation to process variations and improved decision-making in real time. The developed control strategy was implemented and analyzed in the MATLAB/Simulink environment under different operating scenarios, including parameter uncertainties, reference signal changes, and external disturbances. System performance was evaluated using settling time, overshoot, tracking accuracy, steady-state error, and disturbance rejection capability. Simulation results demonstrated that the proposed controller significantly improved dynamic response characteristics, enhanced tracking performance, and increased robustness compared with conventional control approaches. Furthermore, the adaptive learning capability of ANFIS enabled effective operation without frequent controller retuning. The proposed ANFIS-based adaptive predictive control framework provides an efficient and reliable solution for intelligent control of complex nonlinear systems and supports the implementation of advanced automation technologies in Industry 4.0 applications.