TheoremBase

The Extremal Values of tr⁡(AX)\operatorname{tr}(AX) over Symmetric Matrices between aIaI and bIbI

Over the symmetric matrices A with aI ⪯ A ⪯ bI, the linear function A ↦ tr(AX) attains its maximum b·(sum of positive eigenvalues of X) + a·(sum of negative ones) and its minimum with a and b interchanged.

Statement

In the setting of Symmetric and Positive Semidefinite Matrices, the Operator Norm and Spectral Notation, let n∈Nn\in\mathbb{N}, let a,b∈Ra,b\in\mathbb{R} with a≤ba\le b, let

Aa,b={A∈S(n):aIn⪯A⪯bIn},\mathcal{A}_{a,b}=\{A\in\mathcal{S}(n):aI_{n}\preceq A\preceq bI_{n}\},

and let X∈S(n)X\in\mathcal{S}(n). Write tr⁡\operatorname{tr} for the trace and λ1(X)≥⋯≥λn(X)\lambda_{1}(X)\ge\dots\ge\lambda_{n}(X) for the ordered eigenvalues of XX.

Maximum. There is A∗∈Aa,bA^{*}\in\mathcal{A}_{a,b} such that

tr⁡(AX)≤tr⁡(A∗X)=b∑k∈[n], λk(X)>0λk(X)+a∑k∈[n], λk(X)<0λk(X)for every A∈Aa,b.\operatorname{tr}(AX)\le\operatorname{tr}(A^{*}X)=b\sum_{k\in[n],\ \lambda_{k}(X)>0}\lambda_{k}(X)+a\sum_{k\in[n],\ \lambda_{k}(X)<0}\lambda_{k}(X)\qquad\text{for every }A\in\mathcal{A}_{a,b}.

Minimum. There is A∗∈Aa,bA_{*}\in\mathcal{A}_{a,b} such that

tr⁡(AX)≥tr⁡(A∗X)=a∑k∈[n], λk(X)>0λk(X)+b∑k∈[n], λk(X)<0λk(X)for every A∈Aa,b.\operatorname{tr}(AX)\ge\operatorname{tr}(A_{*}X)=a\sum_{k\in[n],\ \lambda_{k}(X)>0}\lambda_{k}(X)+b\sum_{k\in[n],\ \lambda_{k}(X)<0}\lambda_{k}(X)\qquad\text{for every }A\in\mathcal{A}_{a,b}.

Proofs

Log in to submit a proof.

Loading...

Citations

Loading…

Dependencies

Loading…

Related

0 relations

Curated associations between results. These are editable and subjective — they do not replace the dependency graph, which is derived from the references in the text.

No relations recorded yet.

Comments

Log in to comment.

Loading…