A DLQR designed by means of a genetic algorithm for DC-DC boost converters

In this work, a genetic algorithm is employed to design a discrete linear quadratic regulator (DLQR) for a boost converter subject to switched loads. The model of the converter takes into account a parasitic resistance, a delay from digital control signal implementation and an integral action over the error. A genetic algorithm is used to search and provide an optimal choice for the weighting matrices of the DLQR. The objective function of the algorithm is evaluated using data from a standard circuit simulator, allowing to easily include nonlinearities as quantization and saturation. The control law is implemented using a digital signal processor leading to goods experimental results, which are also superior to those from a conventionally designed DLQR.