LQ evolution algorithm optimizer for model predictive control at model uncertainty

This paper presents an evolution algorithm as a powerful optimisation technique for tuning Model Based Predictive Control (MBPC) at the implications of different levels of model uncertainties. Although Standard Genetic Algorithms (SGAs) are proven to successfully tune and optimise MBPC parameters when no model mismatch. SGAs are trapped in a local optimum at the price of model uncertainty. The multi-objective evaluation algorithms are capable to incorporate many objective functions that can meet simultaneously robust control design objective functions. These promising techniques are successfully implemented to stabilised MBPC at high model uncertainty.