Formula 1 engineers have long explored ways to find the optimal racing line to aid the drivers in the quest for ultimate performance. This project aimed to leverage optimization algorithms learnt in the Numerical methods class to develop an algorithm to find the optimal racing line. This was a group project and we used vanilla Gradient Descent (GD) optimization, GD with Momentum and GD with ADAM to find the optimal racing line through a 90-degree corner using the Python Library PyTorch. We also used the project to highlight the differences between the different optimization algorithms.

I am very passionate about Formula 1 and I wanted to leverage the numerical methods I learned in class and apply them to Formula 1. The literature review revealed to me that using Newtonian physics and geometry I formulated a simple problem by discretizing the domain, that would allow us to use different optimization techniques. I found the geometric equation that gave us the instantaneous radius of curvature through the domain. I also contributed to the code, I wrote the vanilla gradient descent algorithm.