Predictive control is a powerful alternative to traditional control, especially when handling complex systems with unknown dynamics and parameters. Predictive control can make use of multivariable input, infer the underlying model through observations and use predictions of future state in a control loop.
Predictive control can in many systems be self-configuring, saving time and cost.
Predictive control is used when there is a need for step by step optimization, common in systems with built in inertia, for example heating and cooling systems. The optimization is done by finding optimal control settings using a predictive model in combination with an action optimizer. Objectives for the optimizer could be reaching the desired temperature, avoid sudden temperature changes and energy efficiency.
Using predictive models in combination with incremental learning (real-time learning on streaming data), the models can learn each part in the system individually, for example learning the temperature dynamics of each room in a building, and thereby finding each part’s individual settings for best performance. Having each part of the system optimized, the system as a whole will improve, typically consuming less energy and optimizing the number of machinery units needed.
When producing and applying system settings the predictive control supports self-configuration. Self-configuration reduces the number of manual steps when installing and maintaining systems, saving time and costs.