Bourbaki

Remark

Tag 083E, permanent. Cite it as [Bourbaki, Tag 083E].

Functions of a Real Variable, chapter I, § 1. FIRST DERIVATIVE

If U(x)U(x) is a square matrix which is differentiable and invertible at the point x0x_0, then the derivative of its determinant Δ(x)=det⁡(U(x))\Delta(x) = \det(U(x)) can be expressed through the derivative of U(x)U(x) by the formula

Δ′(x0)=Δ(x0)⋅Tr⁡(U′(x0)U−1(x0)). \Delta'(x_0) = \Delta(x_0) \cdot \operatorname{Tr}(U'(x_0) U^{-1}(x_0)).
Indeed, let us put U(x0+h)=U(x0)+hVU(x_0 + h) = U(x_0) + h V; then, by definition, VV tends to U′(x0)U'(x_0) when hh tends to 0. One can write
Δ(x0+h)=Δ(x0)⋅det⁡(I+hVU−1(x0)). \Delta(x_0 + h) = \Delta(x_0) \cdot \det(I + h V U^{-1}(x_0)).
Now det⁡(I+hX)=1+hTr⁡(X)+∑k=2nλkhk\det(I + h X) = 1 + h \operatorname{Tr}(X) + \sum_{k=2}^n \lambda_k h^k, the λk\lambda_k (k≥2k \geq 2) being polynomials in the elements of the matrix XX; since the elements of VU−1(x0)V U^{-1}(x_0) have a limit when hh tends to 0, we indeed obtain the formula (3).

Bourbaki puts the proof directly after the statement, so what is above is the statement and its proof together.

Cited by

Nothing.

Cites

Nothing.

Transcribed from the printed volume. Not translated.