Lecture 20: Inequality-constrained Quadratic Programming - Example

In this lecture we work through an example using the active set IQP algorithm derived in Lecture 19

[haroonfaisal]

Thank you for the video !
Really clear walk through of the iterations !!

[matthewjames]

Thanks for the video! This was great! May I ask if there is a way to get a closed form solution to the IQP without iteration, though? Perhaps by using a step function or if statement, you can write into the closed form expression which constraints are active?

[BananenBrot]

Really good stuff. Had to watch a couple of times (mainly lecture 19). One question tho, how would you modify the active set algorithm to deal with additional linear equality contraints? Perhaps consider them always part of the active set?

[eduardosr]

It is a really helpful video, thank you! Just one quick question: In minute 27:00, after you calculate x(3) = [4 1], I understand the meaning of that point, but, why is needed? I mean, we can just verify which constrains are violated by the point x(2) and then add those violated constrains as part working set of the EQP right?
Again, awesome video :)

[sdaniel]

Is "t" always between 0 and 1?

Best

Shawn, grad student

[paugasolina]

suppose we violated more than 1 constraint after solving the KKT system. do we have to add all of the violated constraints to the active set or just one?  Also why is the active set right at the beginning only {1, 2} shouldnt 3 be included aswel? I see that 1, 2 imply 3 but from an algorithmical poin of view 3 should be included aswell?

[francisb.lavoie]

Why the q vector is not updated at each iteration?