I have the following norm in $C^n[a,b]$:
$$\| f \| = \sum_{k=0}^{n-1}|f^{(k)}(a)| + \sup_{[a,b]} |f^{(n)}|$$
I want to show that:
$$(\lambda_0,...,\lambda_{n-1},\mu)\mapsto \left( f\mapsto\sum_{k=0}^{n-1}\lambda_kf^{(k)}(a) + \int_a^b f^{(n)}(t)d\mu(t)\right)$$
Is an isometric isomorphism from $\mathbb{C}^n\oplus M([a,b])$ onto $(C^n[a,b])^*$, with the norm:
$$\|(\lambda_0,...,\lambda_{n-1},\mu \| = \max \{|\lambda_0|,...,\lambda_{n-1},\|\mu \|\}$$
Where $M([a,b])$ is the space of complex Lebesgue-Stieltjes measures on $[a,b]$, and $\| \mu \|$ is the total variation of $\mu$.
I think that it's clear that the mapping is surjective, but I don't know how to show that it's an isometry. At first, I thought to show that the functional linear described is bounded, But I don't know how this can help me.
I do however know that the norm:
$$\| f\| = \max_{k=0,...,n} \sup_{[a,b]}|f^{(k)}|$$
is an equivalent norm on $C^n[a,b]$, but I don't know if I need to use that to show the isometry.
I'm struggling with this question too much, I could use a hint to figure out what I'm supposed to do. Thank you for the help!
2026-03-26 11:04:29.1774523069
Isometric Isomorphism From $\mathbb{C}^n \oplus M([a,b])$ Onto $(C^n[a,b])^*$
99 Views Asked by Bumbble Comm https://math.techqa.club/user/bumbble-comm/detail At
1
First, let's go through the same process as you would use to show that the map of interest is bounded. I'll call it $T$, and I'll write $\lambda := (\lambda_0, \cdots, \lambda_{n-1})$.
We have: $$\begin{split} \left\|T(\lambda,\mu)\right\| &= \sup_{\|f\| \leq 1} \left|T(\lambda, \mu)(f)\right|\\ &= \sup_{\|f\| \leq 1} \overbrace{\left|\sum_{k=0}^{n-1}\lambda_kf^{(k)}(a) + \int_a^b f^{(n)}(t)\mathrm{d}\mu(t)\right|}^{=:\, G_4(f)}\\ &\leq \sup_{\|f\| \leq 1} \overbrace{\left(\sum_{k=0}^{n-1}\left|\lambda_kf^{(k)}(a)\right| + \int_a^b \left|f^{(n)}(t)\right|\mathrm{d}|\mu|(t)\right)}^{=:\,G_3(f)}\\ &\leq \sup_{\|f\| \leq 1} \overbrace{\left(\|\lambda\|_\infty\sum_{k=0}^{n-1}\left|f^{(k)}(a)\right| + \sup_{x \in [a,b]} \left|f^{(n)}(x)\right| \underbrace{\int_a^b \mathrm{d}|\mu|(t)}_{= \|\mu\|}\right)}^{=:\,G_2(f)}\\ &\leq \sup_{\|f\| \leq 1} \overbrace{\max(\|\lambda\|_\infty, \|\mu\|)\|f\|}^{=:\, G_1(f)} = \|(\lambda, \mu)\|\end{split}$$
Now, since we want to prove that $T$ is an isometry, we'll a fortiori have that all the inequalities above are actually equalities.
Conversely, and most interesting to us, if they are equalities then we'll obtain that $T$ is an isometry, thus this gives us a good opportunity to reach our goal. Not only that, but here it suffices to show that a single well-chosen function $f$ maximises all the $\sup$s.
Indeed, assume a function $f \in \mathcal{C}^n(a,b)$ of norm $1$ is built such that: $$\|(\lambda, \mu)\| = G_1(f) = G_2(f) = G_3(f) = G_4(f) =: K$$ From our previous endeavour, we know that: $$\|T(\lambda, \mu)\| = \sup_{\|h\| \leq 1} G_4(h) \leq \sup_{\|h\| \leq 1} G_3(h) \leq \sup_{\|h\| \leq 1} G_2(h) \leq \sup_{\|h\| \leq 1} G_1(h) \leq K$$ Thus, by a finite induction argument using that $G(f) = K \geq \sup_h G(h)$ implies that $K = \max_h G(h)$: $$\|T(\lambda, \mu)\| = \|(\lambda, \mu)\|$$ and $T$ will be an isometry.
You are able to define a good function $f$ for all the cases below:
This lets you demonstrate that $T$ is indeed an isometry, and thus an isometric isomorphism by surjectivity (I don't know if surjectivity is that obvious but if it is for you then you'll be done).