Parameter estimation of a heating system using a genetic algorithm

Presents a genetic algorithm (GA)-based method for the online parameter estimation of a heating system. A more commonly used recursive least square (RLS) estimator is compared with the GA estimator. Both the methods are used with the generalised predictive controller (GPC), to provide self-tuning control to the heating system. Results show that the GA estimates are accurate, its performance is not changed even when the GPC parameters are altered, while the RLS estimator provided some poor estimates and resulted in poor initial control performance by the GPC