Integrating Particle Swarm Optimization with Analytical Nonlinear Model Predictive Control for nonlinear hybrid systems

The computation load remains the main challenge facing the control techniques of hybrid systems with discrete and continuous control signals. In this paper, a new hybrid controller based on Analytical Nonlinear Model Predictive Control (ANMPC) and Particle Swarm Optimization (PSO) for nonlinear hybrid systems is presented. The proposed controller offer sub-optimal solution in reasonable time while respecting the given constraints. The new developed technique is not considered as a computation burden, thus real-time implementation is possible for many hybrid systems. Besides, it can be applied directly to the nonlinear models, avoiding linearization which may lead to inaccurate model and unexpected behaviour. An application of the proposed controller to a three tanks example is presented.