Stochastic numerical treatment for thin film flow of third grade fluid using unsupervised neural networks

In the present study, novel soft computing techniques are developed for numerical treatment of non-linear thin film flow (TFF) problem of third grade fluids using artificial neural networks (ANNs), particle swarm optimization (PSO), sequential quadratic programming (SQP), and their hybrid combinations. The strength of universal function approximation capabilities of ANNs is exploited in formulation of mathematical model of the problem based on an unsupervised error. The training of the design parameter of the networks is performed with PSO, SQP, and hybrid approach PSO-SQP. The proposed schemes are evaluated on four variants of the two cases of TFF problems by taking different values of material parameter and Stokes number. The reliability and effectiveness of the proposed approaches are validated through the results of statistical analyses based on sufficient large number of independent runs.