Let $[a,b]$ an interval. Let $\Sigma$ the set of the subdivision of $[a,b]$. If $$\sigma :a=x_0<x_1<...<x_n=b$$ is a subdivision of $[a,b]$, we denote $$S_\sigma =\sum_{i=1}^n m_i(x_{i}-x_{i-1})\quad \text{and}\quad S^{\sigma }=\sum_{i=1}^n M_i(x_{i}-x_{i-1}),$$ where $$m_i=\min_{[x_{i-1},x_i]}f\quad \text{and}\quad M_i=\max_{[x_{i-1},x_i]}f.$$ I denote $$\overline{S}=\inf_{\sigma \in \Sigma} S^\sigma \quad \text{and}\quad \underline{S}=\sup_{\sigma \in \Sigma}S_\sigma .$$

Question 1 : Fo example, by definition of $\overline{S}$, can I say that there is a sequence $(\sigma _n)_{n\in\mathbb N}$ of $\Sigma$ s.t $$\overline{S}=\lim_{n\to \infty }S^{\sigma _n}\ \ ?$$

Question 2 : I have truble with the sequence $(\sigma _n)_{n\in\mathbb N}$. Can I say that such a sequence is decreasing in the sense that $\sigma _{n+1}\subset \sigma _n$ ? Or can I says that $(\sigma _n)_{n\in\mathbb N}$ is s.t. $h^{\sigma _{n+1}}\leq h^{\sigma _n}$ ? I recall that if $$\sigma _n : a=x_0<x_1<...< x_{n_1}=b$$ and $$\sigma _{n+1} : a=y_0<y_1<...<y_{n_2}=b,$$ then $\sigma _{n+1}\subset \sigma _n$ mean that for all $i=0,...,n_2-1$, there is $j\in \{0,...,n_1-1\}$ s.t. $[y_i,y_{i+1}]\subset [x_j,x_{j+1}]$ . And I denote $h^{\sigma _n}=\max\{|x_{i+1}-x_i|\ :\ i=0,...,n_{1}-1\}$ and same for $h^{\sigma _{n+1}}$.

Question 3 : With the definition inclusion for subdivision, can I say that $(\sigma \cup\tau)\subset \sigma $ and $(\sigma \cup\tau)\subset \tau$ ? Normally, for set it's the converse, but with subdivision, $\sigma \cup \tau$ looks to be the biggest subdivision $\kappa$ s.t. $\kappa \subset \sigma $ and $\kappa\subset \tau$. By the way, what would be $\sigma \cap \tau$ if it exist ?

Question 4 : Can I say that $f$ is Riemann integrable if and only if it's there is a sequence of step function $(f_n)$ s.t. $f_n(x)\to f(x)$ pointwise ? From my definition, the fact that Riemann integrable function are pointwise limit of step function looks straitforward. Is the converse true ?

Question 5: If Q4) is true, does it mean that $$\mathcal R[a,b]=Closure\{step\ function\}\ \ ?$$ where $\mathcal R[a,b]$ is the set of Riemann integrable function over $[a,b]$.

I'm sure that my questions are obvious for all of you, but I always have confusion with all theses things, so I would like to clarify all that.


Q1) Yes of course.

Q2) Yes you can suppose that $h^{\sigma _n}\to 0$.

Q4) Notice that you can define $$\underline{S}=\sup\left\{\int_a^b \varphi\mid \varphi\leq f,\ \varphi\ step\ function\right\},$$ $$\overline{S}=\inf\left\{\int_a^b \varphi\mid f\leq \varphi,\ \varphi\ step\ function\right\},$$ but if $f\in R[a,b]$ you won't have necessarily a sequence of step function $(f_n)$ s.t. $f_n(x)\to f(x)$ for all $x$ (it will be only almost everywhere).

Q5) Closure in which sense ? in $L^1$ ? Unfortunately, you will only have $$R[a,b]\subset Closure\{step\ function\}$$ since there are sequence of step function that converge to function that are not Riemann integrable (for example, there is a sequence of step function that will converge to $\boldsymbol 1_{\mathbb Q\cap [a,b]}$ that is not Riemann integrable.)


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.