Twice Differentiability of a Sum in Separated Variables
lemmaAnalysisMultivariable Calculuslem:twice-differentiable-separated-2026aA sum in separated variables is twice differentiable at a point exactly when both summands are, and its Hessian is then the block diagonal matrix built from their Hessians; in particular the off-diagonal block vanishes.
We work in the setting of Real Matrices, Symmetric Matrices and the Semidefinite Ordering: Standing Notation, whose notation is fixed for every dimension and is used here with natural numbers and satisfying and ; we put . The real numbers, the absolute value , the initial segments , and Euclidean space with its sum and difference of points, scalar multiples, dot product, Euclidean norm , distance , notion of openness and origins , the real matrices and the matrix-vector product , the sets of symmetric real matrices and the concatenation map , a bijection, are all as fixed there. For and the block diagonal matrix , and for the blocks , and , are those of Block Diagonal Symmetric Matrices and the Blocks of a Symmetric Matrix §diagonal and Block Diagonal Symmetric Matrices and the Blocks of a Symmetric Matrix §blocks. Twice differentiability at a point with a given first-order coefficient and Hessian is as defined there.
Let and be open, put
let and , and let be the function determined by
which is well defined because is injective, so that each point of arises from exactly one such pair. Let and , and put . Then the following hold.
1. (The product is open) ¶ is open in .
2. (Sums of twice differentiable functions) ¶ Suppose is twice differentiable at with first-order coefficient and Hessian , and is twice differentiable at with first-order coefficient and Hessian . Then is twice differentiable at with first-order coefficient and Hessian .
3. (Splitting of a twice differentiable sum) ¶ Conversely, suppose is twice differentiable at with first-order coefficient and Hessian , and let be the unique pair with . Then is twice differentiable at with first-order coefficient and Hessian , the function is twice differentiable at with first-order coefficient and Hessian , every entry of equals , and
Loading…
Prerequisites
No prerequisites tracked.
Dependents
No dependents yet.
Dependent proofs
No dependent proofs yet.
No relations recorded yet.