Let $f:[0,T]\rightarrow \mathbb{R}$ $$V(f) = \lim_{\|\Pi\|\rightarrow 0}\sum_{j=0}^{n-1}|f(t_{j+1})-f(t_j)|$$

where $\Pi$ is a partition of $[0,T]$ and define $\|\Pi\|$ as

$$\Pi=\{t_0,t_1,\cdots,t_n\}, \ \ 0=t_0<t_1<\ldots<t_n=T, \ \ \|\Pi\|=\max_i(t_{i+1}-t_i)$$

Now the solution manual says that

Suppose $V(f)$ is finite. Then for any $\epsilon>0$, there exists an $N\geq 1$, $$\sum_{j=0}^{n-1}|f(t_{j+1})-f(t_j)|<V(f) + \epsilon$$ for all $n\geq N$.

I am confused about the description of $+\epsilon$ and $\forall n \geq N$ parts. How do they and this description imply finite of $V(f)$?

I believe this is a trick widely used in real analysis; however, cannot still understand this.


It just follows from the definition of $V(f)$ as a limit.

Given $\epsilon>0,$ there is a $\delta>0$ such that for $any$ paritition $P=\{0,x_1,\cdots ,x_{n-2},T\}$ of mesh less than $\delta$, we have

$\left|\sum_{j=0}^{n-1} |f(t_{j+1}) - f(t_j)|-V(f)\right|<\epsilon\tag1$

So, choose a partition $P$ with $N-1$ large enough so that the mesh of $P$ is less than $\delta.$ Then, $(1)$ holds, from which it follows that

$-\epsilon+V(f)<\sum_{j=0}^{N-1} |f(t_{j+1}) - f(t_j)|<V(f)+\epsilon\tag2$

But any partition with $n>N$ that refines $P$ will have mesh less than $\delta$ so in fact $(2)$ holds for $\textit{all}\ n\ge N$ and this is exactly the result you have in blockquotes.


If $a_n$ is a sequence whose limit exists (call it $a := \lim_{n \to \infty} a_n$), then by the definition of a limit, for each $\epsilon > 0$ there exists $N$ such that $|a_n - a| < \epsilon$ for all $n \ge N$. Note that $|a_n - a| < \epsilon$ implies $a_n < a + \epsilon$.

Your situation is pretty much the same, except the limit is taken as $\|\Pi\| \to 0$. So for any $\epsilon > 0$, there exists a $\delta$ such that $$\sum_{j=0}^{n-1} |f(t_{j+1}) - f(t_j)| < V(f) + \epsilon$$ for any partition satisfying $\|\Pi\| < \delta$. Presumably you have omitted some condition (e.g., you are considering evenly-spaced partitions) that would imply that partitions $\Pi$ of size $\ge N$ will have small $\|\Pi\|$.


What are you asking for the converse of what is sated in the manual. The manual says that if $V(f)<\infty$ the something happens. The stated property is immediate from the definition of limit. It is something like this: if $\lim a_n=l$ and $\epsilon >0$ then there exists $N $ such that $a_n <l+\epsilon$ for $n \geq N$. Do you agree with this statement?


Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

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