A Proof of the Change of Variable Formula for d-Dimensional Integrals

From MaRDI portal





Let \(\lambda\) denote Lebesgue measure in \(\mathbb{R}^d\), \(GL(d,\mathbb{R})\) the group of invertible \(\mathbb{R}\)-linear transformations of \(\mathbb{R}^d\), \(SL(d,\mathbb{R})\) its subgroup of transformations having determinant 1. For each \(A\in GL(d,\mathbb{R})\), \(\lambda\circ A^{-1}\) is a translation-invariant Borel measure, so by the uniqueness of Lebesgue measure it is a nonnegative multiple \(h(A)\) of \(\lambda\). The authors present a novel proof that \(h(A)=|\text{det}(A)|^{-1}\). Besides the cited uniqueness property of \(\lambda\), it uses the facts (I) \(SL(d,\mathbb{R})\) is the commutator subgroup of \(GL(d,\mathbb{R})\), and (II) every continuous endomorphism of the group \((\mathbb{R}^+,\cdot)\) has the form \(\mu(x)= x^c\) for some \(c\in\mathbb{R}\). It is easy to show that \(h\) is a homomorphism into \((\mathbb{R}^+,\cdot)\), so by (I) it factors as \(h= \mu\circ\text{det}\). Using \(A\) whose \((d-1)\)-by-\((d-1)\) southeast corner is the identity matrix \(E_{d- 1}\), it is shown that det is an open mapping, whence continuity of \(\mu\) follows from that of \(h\). Therefore, \(\mu\) has the form prescribed in (II). The exponent \(c\) is identified as \(-1\) by examining \(A:= 2E_d\) and the effect of \(\lambda\) and \(\lambda\circ A^{-1}\) on the unit cube.











This page was built for publication: A Proof of the Change of Variable Formula for d-Dimensional Integrals

Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q4222455)