Artificial intelligence tools (such as evolutionary algorithms, and fuzzy logic algorithms) are useful for solving the more complex problems that arise in identification, modelling, and process control. These tools can help answer the following questions:
How to optimally tune a multivariable predictive controller?
How to adjust the parameters of a proportional–integral–derivative controller (PID controller) with respect to multiple criteria (stability, robustness, performance, and implementation)?
How to find the worst case in a simulation to demonstrate that a control system is safe?
How to use plant data to discover parameters in a model based on first principles?
How to optimally adjust the parameters of an inferential sensor?
Can an evolution optimiser be used to control a process in real time?
In this line of research, the CPOH group is exploring the feasibility and application of artificial intelligence tools to solve problems in systems engineering and automation. These tools are intended to be applied in the field of process control (such as advanced control, and optimisation of setpoints); as well as other engineering areas related to modelling and identification of systems. The algorithms and techniques used enable more robust, manageable, and cheaper solutions to be obtained than those produced by conventional techniques. When the computational cost of the problem is high, a wide and heterogeneous variety of problems can be approached using evolutionary algorithms and parallel computing techniques.