Convex optimization
- Instructor: Alex L. Wang, Krannert 463, wang5984@purdue.edu
- Lectures: TuTh 3:30-5pm RAWL 2077
- Gradescope: Link
- Piazza: Link
- Office hours: By appointment
- Notes: Draft - These notes will be updated throughout the semester.
I apologize for many missing references. I will clean up this document and add citations soon.
Please email me regarding any errors. - Optional texts:
- LMCO: Lectures on Modern Convex Optimization, Ben-Tal and Nemirovski
- LCO: Lectures on Convex Optimization, Nesterov
- CC: A Course in Convexity, Barvinok
- FOMO: First-Order Methods in Optimization, Beck
Course policies can be found here
Schedule
Date |
Topics |
Optional reading |
Linear algebra and convex analysis background | ||
3/5 |
Lec 1 Linear algebra review |
LMCO Appendix A |
3/7 |
Lec 2 Elementary convex analysis I |
CC 1.1–1.2, 2.1–2.8, 2.12 |
3/12 |
Spring break no classes |
|
3/14 |
Spring break no classes |
|
3/19 |
Lec 3 Elementary convex analysis II |
CC 1.1–1.2, 2.1–2.8, 2.12 |
Conic programming | ||
3/21 |
Lec 4 Conic programming I |
LMCO 1.1, 1.2, 1.4 |
3/24 |
||
3/26 |
Lec 5 Conic programming II |
LMCO 1.4 |
Conic programming applications | ||
3/28 |
Lec 6 SOCP representability |
LMCO 2.1–2.3 |
4/2 |
Lec 7 SDP representability |
LMCO 3.1, 3.2 |
4/4 |
Lec 8 SDP relaxations of intractable problems |
CC 4.10–4.11, LMCO 3.3–3.4 |
4/7 |
||
First-order methods for smooth minimization | ||
4/9 |
Lec 9 Subgradient descent, gradient descent, accelerated gradient descent |
LCO 3.1.5, 3.2.3 |
4/11 |
Lec 10 Subgradient descent, gradient descent, accelerated gradient descent |
LCO 2.1.1, 2.1.3, 2.1.5, 2.2 |
4/16 |
Lec 11 Oracle lower bounds |
LCO 3.2.1, 2.1.2, 2.1.4 |
4/18 |
Lec 12 Convex conjugates and performance estimation programming |
FOMO 4 |
4/21 |
||
Other topics in first-order methods | ||
4/23 |
Lec 13 Mirror descent |
FOMO 9 |
4/25 |
Lec 14 Frank–Wolfe Method |
[Sebastian Pokutta’s blog] On Frank–Wolfe and extensions (see posts 2–4) |
5/2 |