Is it ok to solve min max problem by cvx?

My problem is formulated as
min max_k {f_k(x)}
s.t. x>=0;
where f_k, \forall k, is convex.

Is it ok to replay max_k{f_k(x)} with v, then the problem is converted into

min v
s.t. v>=f_k(x), \forall k,

Sure, that’s just fine! That’s not just equivalent in CVX, but it’s equivalent mathematically.

1 Like