Why doesn’t the Gekko solver adapt to variations in the system? The following is related to this question : predictive control model using GEKKO