Application of optimal RBF neural networks for optimization and characterization of porous materials

A hybrid MPSO-BP structure adaptive algorithm for RBFNs

“…An increasing amount of literature is available on RBFNs methods which became a popular technique since the 1980s because of their simple structure, well established theoretical basis and fast learning speed, which are all crucial factors in real applications [1][2][3][4][5]. However, there are still some difficulties with building RBFNs.…”

Application of the LBET class models to describe the structure of microporous activated carbons on the basis of argon and benzene adsorption isotherms

“…Several different equations have been proposed to describe the adsorption equilibrium, among which the most popular are: the Dubinin-Radushkevich equation (Dubinin 1960(Dubinin , 1989, the Dubinin-Astakhov equation (Dubinin and Astakhov 1971;Gil and Grange 1996), and the BET equation (Gregg and Sing 1967;Gomez-Serrano et al 2001). The development of computer science and the improvements in computational technology have triggered the development of more advanced methods of pore structure description (Rudziński and Everett 1992;Puziy et al 1997), based on sophisticated numerical tools (Puziy et al 1997), statistical mechanics (Sánchez-Montero et al 2005;Cao et al 2002), computer simulations (Nicholson 1994;Suzuki et al 1996), the DFT theory (Ryu et al 1999), fractal geometry (ErdemSenatalar et al 2000) and neural networks (Shahsavand and Ahmadpour 2005). Although ranges of advanced methods are available, the obtained results are still unsatisfactory.…”

Artificial Neural Network for Cooperative Distributed Environments

“…This strategy combines Neural-Gas (NGAS) (Martinetz & Schulten, 1991), Radial Based Function Network (RBFN) (Lingireddy & Ormsbee, 1998) (Shahsavand & Ahmadpour, 2005) and Multi-Layer Perceptron (MLP) (Haykin, 1998) ANN-based models aiming to cover different aspects in the learning capabilities of AMBAR: 1) a supervised-based method for learning for learning how to collaborate based on levels of awareness; 2) an unsupervised-based method for selecting a potential candidate to negotiate on saturated conditions; and 3) a supervised-based method to learn the decision whether or not a node must change the information that describes its current conditions related with collaboration. Just as a quick reminder, NGAS is a Vector Quantization (VQ) (Kohonen et al, 1984) (Makhoul et al, 1985) (Nasrabadi and Feng, 1988) (Nasrabadi & King, 1988) (Naylor & Li, 1988) technique with soft competition between the units.…”