Take the 2-minute tour ×
Mathematics Stack Exchange is a question and answer site for people studying math at any level and professionals in related fields. It's 100% free, no registration required.

If $N(\bullet)$ is a matrix norm, it is unitarily invariant if and only if for any unitary matrices $U$, $V$, and for any matrix $A$ it satisfies: $$N(A)=N(UAV).$$

To say that they form a convex set, is to say that for any unitarily invariant matrix norm $N$, $M$, and for any $0\le \lambda\le 1$,$$P(A)=\lambda N(A)+(1-\lambda)M(A)$$ is a unitarily invariant norm.

It's easy to show that $P(A)\ge0$, $P(A)=0$ if and only if $A=0$, $P(cA)=|c|P(A)$ and $P(A+B)\le P(A)+P(B)$. The only property yet to prove is sub-multiplicative: $$P(AB)\le P(A)P(B).$$

To prove this, I notice that $$\begin{align}P(AB)&=\lambda N(AB)+(1-\lambda)M(AB)\\&\le\lambda N(A)N(B)+(1-\lambda)M(A)M(B)\end{align}$$and that $$\begin{align}P(A)P(B)&=\bigg(\lambda N(A)+(1-\lambda)M(A)\bigg)\bigg(\lambda N(B)+(1-\lambda)M(B)\bigg)\end{align}$$

but is it possible to prove that $$\lambda N(A)N(B)+(1-\lambda)M(A)M(B)\le \bigg(\lambda N(A)+(1-\lambda)M(A)\bigg)\bigg(\lambda N(B)+(1-\lambda)M(B)\bigg)$$?

share|improve this question
add comment

Your Answer

 
discard

By posting your answer, you agree to the privacy policy and terms of service.

Browse other questions tagged or ask your own question.