OptimLi (inactive)

A multidimensional optimization code

In many engineering applications, it is desireable to define an objective function based on design parameters and minimize that objective function. This is common, for example, in structural engineering where it is desireable to keep the Von Mises stress in a structural member below the yield stress; it is also desireable in an aeroplane to design a wing which produces the highest lift-to-drag ratio.

The OptimLi project was started in 2017 to review some of the results obtained in my Capstone project. I later started using it in my research and decided to release the code as open source for easier access. OptimLi is designed to solve simple optimization problems in a domain where the number of unknown variables is small. Under the hood, OptimLi is using LAPACK and BLAS, which are not themselves designed for highly scalable systems. The purpose of OptimLi is to efficiently solve simple optimization problems where a Hessian matrix may be supplied.

Due to my lack of formal programming education at the start of this project, it ended up being a mixture of C and C++ coding styles which was difficult to maintain and certainly not recommended. In 2024, I decided to scrap this project in favour of either fully rewriting it or allowing it to fade into obscurity.