Mathematical optimization deals mainly with the problem of finding a minimum or a maximum of a mathematical function. Frequently, a real-world numerical problem can be expressed as a function minimization problem. Such examples can be found in statistical inference, machine learning, graph theory, and other areas.

Although there are many function minimization algorithms, a generic and universal method does not exist. Therefore, it is important to understand the differences between existing classes of algorithms, their specificities, and their respective use cases. We should also have a good understanding of our problem and our objective function; is it continuous, differentiable, convex, multidimensional, regular, or noisy? Is our problem constrained or unconstrained? Are we seeking local or global minima?

In this recipe, we will demonstrate a few usage examples of the function minimization algorithms implemented in SciPy.