Mathematics Stack Exchange is a question and answer site for people studying math at any level and professionals in related fields. Join them; it only takes a minute:

Sign up
Here's how it works:
  1. Anybody can ask a question
  2. Anybody can answer
  3. The best answers are voted up and rise to the top

Can someone explain how to differentiate something like

$$\prod\limits_{i<j}^N {(x_i-x_j)}$$

with respect to $x_i$

The product starts from 1 and goes to N. I started off by ignoring the $x_j$ as it doesn't depend on $i$ but I have a feeling that's not correct. What confuses me is the $i<j$

I have tried this on Maple but the answer doesn't make sense to me. Could someone help me out? Is there an explicit formula I should use?

share|cite|improve this question
Is $i$ your index variable and is $j$ fixed or does it change? – Cameron Williams Feb 6 '14 at 17:01
They both change so essentially the first few terms are $$(x_1-x_2)(x_2-x_3)\dots$$ – user113905 Feb 6 '14 at 17:05
The notation $\displaystyle\prod_{i<j}$ sometimes means $\displaystyle\prod_{j\;:\;i<j}$ and sometimes means $\displaystyle\prod_{i,j\;:\;i<j}$ and sometimes means $\displaystyle\prod_{i\;:\;i<j}$. When not disambiguated by the context, this should be made explicit. – Michael Hardy Feb 6 '14 at 17:22
Do you think you can edit your post to give an example i.e. the product for $N=3$ or $N=4$ say? I'm not sure whether $N=3$ is $(x_1-x_2)(x_2-x_3)(x_1-x_3)$ for example or just $(x_1-x_2)(x_2-x_3)$. – TooTone Feb 6 '14 at 17:23
If BOTH variables change, i.e. if you mean $\displaystyle\prod_{i,j\;:\;i<j}$, then what would it mean to differentiate it with respect to $x_i$? Which variable is $x_i$? For example, if $i,j$ range from $1$ to $3$, then the product is $(x_2-x_1)(x_3-x_1)(x_3-x_2)$. In the first two factors, $i$ is $1$, and in the last factor $i$ is $2$. Which variable are you differentiating with respect to? – Michael Hardy Feb 6 '14 at 17:42
up vote 2 down vote accepted

Since $i$ and $j$ are both bound variables of the expression, you should at most differentiate with respect to something "new", $x_k$ say. Then use the product rule: Since $$\frac\partial{\partial x_k}(x_i-x_j)=\begin{cases}1&\text{if }k=i\\-1&\text{if }k=j\\0&\text{otherwise}\end{cases}$$ we obtain $$\begin{align} \frac\partial{\partial x_k}\prod_{1\le i<j\le n}(x_i-x_j)&=\sum_{1\le i<j\le n} \frac{\partial(x_i-x_j)}{\partial x_k}\cdot\prod_{1\le \iota<\nu\le n\atop(\iota,\nu)\ne(k,j)}(x_\iota-x_\nu)\\ &=\sum_{k<j\le n} \prod_{1\le \iota<\nu\le n\atop(\iota,\nu)\ne(i,j)}(x_\iota-x_\nu)-\sum_{1\le i<k} \prod_{1\le \iota<\nu\le n\atop(\iota,\nu)\ne(i,k)}(x_\iota-x_\nu)\\ &=\sum_{j=k+1}^n \prod_{1\le \iota<\nu\le n\atop(\iota,\nu)\ne(i,j)}(x_\iota-x_\nu)-\sum_{i=1}^{k-1} \prod_{1\le \iota<\nu\le n\atop(\iota,\nu)\ne(i,k)}(x_\iota-x_\nu).\end{align}$$ It may be easier to try with the logarithmic derivative: $$\frac{\frac\partial{\partial x_k}f}{f}= \frac\partial{\partial x_k}\ln f =\frac\partial{\partial x_k}\sum_{1\le i<j\le n}\ln(x_i-x_j)=\sum_{k<j\le n}\frac1{x_k-x_j}-\sum_{1\le i<k}\frac1{x_i-x_k}$$ although this is formally correct only if $x_1>x_2>\ldots >x_n$.

share|cite|improve this answer
Many thanks for your help! – user113905 Feb 6 '14 at 17:39

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.