<![CDATA[A hybrid non-dominated sorted genetic algorithm II with a random forest. The polynomial regression (HNSGAII-RFPR) method for the photovoltaic thermal system (PTS) optimization problem is proposed. This new method uses polynomial regression to generate an objective function, which at first contains essential and non-essential terms of the parameters for the PTS. The random forest algorithm was used on the regression model to identify the important terms of the parameters by deploying the mean squared error metric, which showed the difference between prediction and experimental values. The parameters used in the process are collector angle, number of collector tubes, volume-area proportion, mass flow rate, collector azimuth, and collector plate material, while the outputs are the electrical and thermal efficiencies. Results showed that the electrical efficiency exhibited acceptable terms, with the mean squared error of 0.09, while the thermal efficiency results were not acceptable with a mean squared error of roughly 130. However, the proposed algorithm performed satisfactorily regarding accuracy. Moreover, experiments with the training and testing data, which are 14/18 and 4/18, respectively, were obtained from the experiment of Kuo et al. The simulation experiments highlight that the use of the HNSGAA II-RFPR schema can produce interesting results for the photovoltaic thermal system. The simulations ensure a good balance between solution convergence and its diversity. The implementation of the tests demonstrated that the proposed method competes with another evolutionary algorithm tested.]]>
Copyrights © 2026
