Instructor

Ankur Kulkarni

Semester

Spring ‘21

Course Difficulty

Course was decent at the beginning and became very difficult towards the end

Time Commitment Required

Apart from class timings, maybe an hour or two in a week for general revision

Grading Policy and Statistics

AA-1, AB-4, BB-12 out of total 55 students

Attendance Policy

None

Pre-requisites

None

Evaluation Scheme

Presentation: 20%
Homeworks/assignments: 40%
Final exam/project: 40%

Topics Covered in the Course

Convex analysis and review of real analysis
Optimization in Euclidean space: linear programming, duality, convex optimization, constrained optimization and KKT conditions, geometric theory of duality (Lagrangian, Fenchel duality)
Algorithms for optimization in Euclidean space
Nonconvex optimization
Discrete optimization, brief overview of complexity theory towards an appreciation of “hard” and “easy” problems.

Teaching Style

In Every class, students in groups of 3 would give a presentation and briefly explain the topics that were covered in the pre-recorded lectures (available on CDEEP). You can choose not to attend and instead spend the time watching the lectures. Doubts were covered in class. No tutorials.

Tutorials/Assignments/Projects

Assignments were multiple choice and usually hard.

Feedback on Exams

Exams were also multiple choice and usually harder than assignments.

Motivation for taking this course

This is a really good course if you like math. It covers a lot of material related to optimization and is widely used in robotics.

How strongly would I recommend this course?

Would recommend it strongly if you enjoy math.

When to take this course?

Fourth semester, as an ALC. Any semester is ideal for this course.

Going Forward

Robotics, structural optimization (Shape and Topology Optimization), Optimal controls

SC 607 Review By: Hiya Gada