Take the 2-minute tour ×
Mathematics Stack Exchange is a question and answer site for people studying math at any level and professionals in related fields. It's 100% free, no registration required.

Let $f$ be a function, that has continuous partial derivatives of order $m+1$ in an open ball $B(a,r)$ in $R^n$. Show for $k=1, \dots, m$ that the equation is valid:

$$\frac1{k!}\sum_{j_1=1}^n \cdots \sum_{j_k=1}^n \frac{\partial^k f}{\partial x_{j_1} \cdots \partial x_{j_k}}(\mathbf a)x_{j_1}\cdots x_{j_k}=\sum_{|\mathbf{\alpha}|=k}\frac1{\alpha!}D_\alpha f(\mathbf a)\mathbf x^\alpha$$

where $\mathbf \alpha$ is an $n$-dimensional multi-index.

share|improve this question

1 Answer 1

When you are told to compute all $k$-th partial derivatives of a function $f$ of $n$ variables $x_j$ then formally there are $n^k$ of them: For each word $(j_1,j_2,\ldots, j_k)\in [n]^k$ you differentiate first with respect to $x_{j_1}$, then with respect to $x_{j_2}$, $\ldots$, and finally with respect to $x_{j_k}$. On the left side of your equation you have the sum of all these $n^k$ derivatives (at the point ${\bf a}$), each multiplied with the corresponding monomial $x_{j_1}\ x_{j_2}\cdots\ x_{j_k}$.

Any two mixed partial derivatives where each variable has been affected the same number of times coincide, and at the same time the corresponding monomials $x_{j_1}\ x_{j_2}\cdots\ x_{j_k}$ coincide. Therefore among the $n^k$ terms on the left of your equation there are a lot of them equal. On the right side equal terms have been collected. The multiindex $\alpha=(\alpha_1,\alpha_2,\ldots, \alpha_n)$ of weight $|\alpha|=k$ encodes how often each of the individual variables $x_i$ is affected: For each $i\in[n]$ you should differentiate $\alpha_i$ times with respect to $x_i$ resp. take the factor $x_i^{\alpha_i}$ into the accompagning monomial. Now it is an elementary combinatorial fact that for each multiindex $\alpha$ of weight $k$ there are exactly $k!/\alpha!$ terms on the left side which correspond to this $\alpha$.

By the way: The large expressions considered here are generated when you develop the auxiliary function

$$\phi(t)\ :=\ f({\bf a} + t{\bf x})\qquad (t\in{\mathbb R})$$ into a Taylor series with respect to the real variable $t$. In order to compute $\phi^{(k)}(t)$ you have to apply the chain rule $k$ times in succession, and each time the formal number of terms is multiplied by $n$.

share|improve this answer

Your Answer


By posting your answer, you agree to the privacy policy and terms of service.

Not the answer you're looking for? Browse other questions tagged or ask your own question.