Kang-Jia Wang
In this work, a novel fractal physics-informed neural network (PINN) method is proposed to solve the fractal exothermic reactions model with a constant heat source and porous media for microgravity. On integrating the physical information of the considered equation into the neural networks with the aid of the two-scale transformation, we convert the numerical approximate solution problem into optimization problems of boundaries and governing equations. The Adam algorithm is adopted to optimize the loss function through the training process to obtain the predicted results. Two different examples are given to validate the robustness and efficiency of the fractal PINNs method. As anticipated, compared with the existing approximate solutions for the fractal order [Formula: see text], a good agreement is reached, which reveals the validity and correctness of the proposed method. Furthermore, the predicted results of the different fractal orders are also presented. The findings in this work are expected to offer some new ideas for solving the fractal nonlinear partial differential equations (NPDEs).