Key points are not available for this paper at this time.
Purpose This paper aims to investigate the effects of thermal radiation on magnetohydrodynamics (MHD) bioconvection nonlinear complex structure flow of non-Newtonian fluids such as Casson, Williamson and Sisko fluids. Design/methodology/approach The nonlinear coupled fundamental equations governing the steady, incompressible combined with Casson–Williamson–Sisko fluids flow over an exponential sheet are reduced to ordinary differential equations using appropriate transformations. Open-source platforms such as Google Colab and Python are used. Results, performance, accuracy and correlation are examined with neural networking, Levenberg-Marquardt, machine learning, artificial intelligence (AI) algorithms and linear regression. Findings Numerical and graphical results are presented to observe the impact of physical parameters. The prospect of AI tools, particularly Levenberg-Marquardt, increases the accuracy of developed complex fluid dynamics models. Besides, the further scope of machine learning in the hybrid nature of fluids is also presented. It is concluded that Levenberg-Marquardt algorithm is the most suitable for the simulation of boundary layer flow with high accuracy, smooth regression curves and the minimum rate of error. It is observed that the range of 10−8 for mean squared error shows the good fit of the model. It is noted that by increasing the Casson and Williamson fluids’ parameters, the velocity profile decreases. Both concentration and motile density decrease with an increasing values of Schmidt and Peclet numbers. Originality/value The existing literature lacks a comparative analysis of neural networks and machine learning in predicting boundary layer flow using AI-based approaches, linear regression algorithm for bioconvection MHD flow of Casson–Williamson–Sisko fluids with thermal radiation in existing literature. This effort is devoted to fill the said gap.
Sait et al. (Wed,) studied this question.