Solving real-world optimization tasks using physics-informed neural computing

Paper link: https://www.nature.com/articles/s41598-023-49977-3

This paper provides 3 examples of using PINNs to solve real-world optimization tasks. The examples are:

1. Swinging up a pendulum
2. Determining the shortest-time path connecting two given points
3. Finding a swingby trajectory of a spacecraft

I have implemented a fourth example: Inverting pendulum attached to cart

The code is avalailible at data/inverting_pendulum/inverted_pendulum_cart.ipynb All the data required for training the model is generated within the notebook itself.

The resource used for the dynamical equations of the system is: Dynamics of inverted pendulum on a cart

More information about the problem and the model is given in the images below: