What is Convex Optimization?

Optimization · Axiom Academy

INTRO What is Convex Optimization? Discover why some optimization problems are surprisingly easy to solve. Step 1: Exploring Different Landscapes Let's start by exploring two different types of functions. Drag the slider to see how they behave. Step 2: The Local vs Global Challenge Click anywhere on the functions below to try finding their minimum values. What challenges do you encounter? Step 3: The Golden Property of Convexity Select points on the convex function and see what happens when we draw a line between them. Step 4: Why This Makes Optimization Easy Watch how different algorithms perform on convex vs non-convex functions. Click "Run Algorithm" to see. Convex optimization is the study of minimizing convex functions over convex sets. It's a special class where local optimality implies global optimality. Any local minimum = Global minimum. This means simple algorithms can reliably find the best solution. Despite being "easy" to solve, convex optimization applies to thousands of real-world problems: machine learning, finance, engineering design, signal processing, and more. The hardest part is often recognizing when a problem is convex, or transforming it into a convex form. Once you do, powerful solvers can handle problems with millions of variables!

This is the written version of the interactive lesson above. See the full Optimization course.