Adam Optimization of Burger’s Equation Using Physics-Informed Neural Networks

In this paper, physics informed neural networks are used for numerical approximation of partial differential equations. The data which is used in the process is generated by Latin Hypercube sampling which has been discussed. Adam optimization technique has been implemented to minimize the loss for the discussed partial differential equation. The above proposed methodology has been applied for Burger’s equation and the obtained results have been discussed in section 5. Loss function graphs have also been provided to showcase the efficiency of the proposed methodologies.