Solubility prediction of supercritical carbon dioxide in 10 polymers using radial basis function artificial neural network based on chaotic self-adaptive particle swarm optimization and K-harmonic means
Abstract
A novel model combined with chaos theory, self-adaptive particle swarm optimization (PSO) algorithm, K-harmonic means (KHM) clustering and radial basis function artificial neural network (RBF ANN) is proposed, hereafter called CSPSO-KHM RBF ANN. Traditional PSO algorithm is modified by chaos theory and self-adaptive inertia weight factor in order to reduce premature convergence problem. The modified PSO algorithm is employed to trim the RBF ANN connection weights and biases, whereas KHM is used to tune the hidden centers and spreads. The CSPSO-KHM RBF ANN model was employed to investigate the solubility of supercritical carbon dioxide in 10 polymers. Compared with other methods, such as RBF ANN, adaptive neuro-fuzzy inference system and PSO ANN, the proposed model displays optimal prediction performance. Results discover that the CSPSO-KHM RBF ANN model is an effective method for solubility prediction with high accuracy, and is a practicable method for chemical process analyzing and designing.