How to let the condition that matrix is nonpositive definite be one of the constraints in CVX

Nonconvex
Apr 19, 2017
X
#1

How to let the condition that matrix is nonpositive definite be one of the constraints in CVX

M
#2

This is a non-convex constraint, except in one dimension.

However, you can constrain a matrix A to be negative semi-definite by constraining its negative to be (positive) semidefinite, or equivalently , by constraining lambda_max(A) <= 0 .

1 reply
X
#3
Replying to #2

Many thanks!Thank you very much!

‘you can constrain a matrix A to be negative semi-definite by constraining its negative to be (positive) semidefinite’.Did you mean that I can denote the constraint as -A==semidefinite(N)?

1 reply
M
#4
Replying to #3

Yes, as per your other question.

1 reply
X
#5
Replying to #4

Ok! Got it! Thank you very much!