|
|
|
|
derivative of homogeneous function
|
(Theorem)
|
|
Proof. By considering component functions if necessary, we can assume that $m=1$ . For
 , let $M_\lambda$ be the multiplication map,
For $\lambda>0$ and
 , we have \begin{eqnarray*} \frac{\partial f}{\partial x^i}(\lambda v) &=&\frac{\partial(f\circ M_\lambda \circ M_{1/\lambda})}{\partial x^i}(\lambda v) \\ &= &\sum_{l=1}^n\frac{\partial(f\circ M_\lambda)}{\partial x^l} (v)\, \frac{ \partial(x\mapsto x/\lambda)^l}{\partial x^i} (\lambda v) \\ &= &\frac{\partial(f\circ M_\lambda)}{\partial x^i} (v)\, \frac{1}{\lambda}\\ &= &\lambda^{r-1}\frac{\partial f}{\partial x^i} (v) \end{eqnarray*}as claimed. 
|
Anyone with an account can edit this entry. Please help improve it!
"derivative of homogeneous function" is owned by matte.
|
|
(view preamble | get metadata)
Cross-references: map, multiplication, necessary, functions, component, positively homogeneous function of degree, differentiable
This is version 6 of derivative of homogeneous function, born on 2004-10-18, modified 2007-09-15.
Object id is 6390, canonical name is DerivativeOfHomogeneousFunction.
Accessed 2600 times total.
Classification:
| AMS MSC: | 15-00 (Linear and multilinear algebra; matrix theory :: General reference works ) |
|
|
|
|
|
|
Pending Errata and Addenda
|
|
|
|
|
|
|
|
|
|
|