Computational intelligence of Levenberg-Marquardt backpropagation neural networks to study thermal radiation and Hall effects on boundary layer flow past a stretching sheet

In the present investigation, an integrated numerical computing through neural networks which are backpropagated with Levenberg-Marquard scheme (NN-BLMS) is designed for the study of thermal radiation and Hall effect on boundary layer flow over a non-isothermal stretching sheet (TRH-BLF-NSS) embedded with non-uniform heat source and fluidic-particle suspension in porous medium. Porosity of the medium, thermal radiation, Hall effect and non-uniform heat source (PTHH) is considered. The sheet is considered workable to permit fluid to suction, and according to a linear varying stretching with a surface velocity. Initially TRH-BLF-NSS was represented by system of PDEs which later converted to a system of non-linear ODEs through appropriate transformation and then solved numerically. At the surface two cases of the temperature boundary conditions namely PST and PHF were considered. By using Adam numerical method, for different scenarios connected with PTHH a dataset is generated for the expected NN-BLMS by variation various parameters. From reference results to perform NN-BLMS training, testing and validation process to find out estimated solutions of PTHH for variants associated with the physical system and to prove the accuracy of proposed NN-BLMS. Through mean square error, regression analysis and histogram the performance of NN-BLMS is studies and then successfully solve PTHH.