2
$\begingroup$

I’m attending a functional analysis course and I am given to solve this problem as an exercise but I’m a little bit disoriented and I don’t know what tools I can use to get it.

Show that, for each $N \in \mathbb{N}$ , there exists a constant $C(N) \in \mathbb{R}^+$ such that if $P$ in a polynomial of degree $N$ with complex coefficients, we have $$\sum_{k=0}^{N} |P(k)| \leq C(N) \int_{0} ^{1} |P(t)| dt\,.$$

Any idea or hint? Thank you all in advance

$\endgroup$
2
$\begingroup$

Let $V_N$ be the complex vector space of polynomials in $\mathbb{C}[t]$ of degree at most $N\in\mathbb{Z}_{\geq 0}$. Show that $\|\_\|_1$ and $\|\_\|_2$ defined by $$\|P\|_1:=\sum_{k=0}^N\,\big|P(k)\big|$$ and $$\|P\|_2:=\int_0^1\,\big|P(t)\big|\,\text{d}t$$ for all $P\in V_N$ are norms on $V_N$.

To show that $\|\_\|_1$ is a norm, we note that $\|a\, P\|_1=|a|\,\|P\|_1$ trivially holds for all $a\in\mathbb{C}$ and $P\in V_N$. If $\|P\|_1=0$ for some $P\in V_N$, then $P(0)=P(1)=P(2)=\ldots=P(N)=0$, so $$P(t)=Q(t)\,\prod_{k=0}^N\,(t-k)$$ for some $Q(t)\in\mathbb{C}[t]$. As $P$ has degree at most $N$, $Q$ must be identically $0$, so $P\equiv 0$. If $P_1,P_2\in V_N$, then $P_1+P_2\in V_N$, and $$(P_1+P_2)(k)=P_1(k)+P_2(k)\text{ for every }k\in\mathbb{C}\,,$$ whence $$\big|(P_1+P_2)(k)\big|=\big|P_1(k)+P_2(k)\big|\leq \big|P_1(k)\big|+\big|P_2(k)\big|\text{ for every }k\in\mathbb{C}\,.$$ Taking the sum over $k=0,1,2,\ldots,N$, we get $\|P_1+P_2\|_1\leq \|P_1\|_1+\|P_2\|_1$, justifying the triangle inequality condition. Similarly, $\|\_\|_2$ is a norm on $V_N$.

Note that all norms on a finite-dimensional complex vector space are equivalent. That means, there exist $c(N),C(N)\in\mathbb{R}_{>0}$ such that $$c(N)\,\|P\|_2\leq \|P\|_1\leq C(N)\,\|P\|_2$$ for all $P\in V_N$.

$\endgroup$
  • $\begingroup$ Thank you! I can observe that the polynomial space I am considering has $N+1$ as dimension because it is generated, for example, by $\{x^{N},x^{N-1},...,x,1\}$ that has dimension $N+1$, which is finite. Right? $\endgroup$ – Maggie94 Dec 10 '18 at 13:03
  • $\begingroup$ Yes, you are right. $\endgroup$ – Batominovski Dec 10 '18 at 20:03

Your Answer

By clicking "Post Your Answer", you acknowledge that you have read our updated terms of service, privacy policy and cookie policy, and that your continued use of the website is subject to these policies.

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