/
lectern.tv
Play15s30s1x0:00/--:--
DoneWatched
Expand
lectern.tv
22:38/25:36
1:26
1:52
1:32
2:03
3:03
2:06
0:58
1:43
1:18
1:15
0:51
4:25
2:24
0:40
The Anatomy of an Optimization Problem
Dissecting the fundamental components of an optimization problem, exploring objective functions, constraints, and the implications of convexity on problem categories.
Lecture guide · Written by AI from the lecture. For the lecture’s own words, open the Transcript tab.
Breaks down the anatomy of an optimization problem, defining objective functions, decision variables, inequality and equality constraints, and the feasible region. It contrasts convex and non-convex problems, explains the roles of gradients and slack variables, and introduces key problem classes such as linear and quadratic programming.
This lecture assumes
Key Takeaways
- An optimization problem consists of an objective function to minimize or maximize, subject to optional inequality and equality constraints.
- Min refers to the optimal objective function value, while argmin refers to the variable argument that achieves that minimum.
- Convex optimization problems have positive bowl-shaped curvature where any local minimum is guaranteed to be a global minimum.
- Constraint equations establish a feasible region, which can be represented as a convex polytope when defined by linear inequalities.
Concepts Explained10
Optimization Problem Anatomy—An optimization problem consists of an objective function to minimize or maximize over optimization variables, subject to optional inequality and equality constraints that define a feasible region.
Min vs Argmin—Min refers to the minimum objective function value itself, whereas argmin refers to the variable argument that achieves that minimum value.
Convex Optimization—Convex optimization involves objective functions with positive curvature shaped like bowls where any local minimum is also a global minimum.
Nonconvex Optimization—Nonconvex optimization involves objective functions with multiple local minima, saddle points, and maxima, making global optimization challenging.
Examples, Proofs & Derivations
Works & Sources Cited1
Paper
Lagemann et al. 2024—Cited as an example of neural network controllers in flow environments.
Part of course·Lecture 2 of 9
ME 565: Optimization Bootcamp: Convex Optimization for ML & Control
About the lecturer
Professor of Mechanical Engineering & Adjunct Professor of Applied Mathematics, University of Washington
University of Washington Professor known for data-driven dynamical systems, SINDy, Koopman operator theory, and control engineering.
2 courses · 22 lectures in archiveView profile
Discussion
Share a seminar note, question, or timestamped reference (e.g. 14:20)...Sign in