1
$\begingroup$

Let $f$ be a continuous function defined on $\mathbb Z_p$. By Mahler theorem, there exists a sequence $(\gamma_k)_{k\in\mathbb N}$ of $\mathbb Z_p$ such that for every $z\in\mathbb Z_p$ $$f(z)=\sum_{k\ge0}\gamma_k\binom{z}{k},$$ where $\binom{z}{k}=\frac1{n!}\prod_{j=0}^{k-1}(z-j)$. I try to prove (without any success) that for every $n\in\mathbb N$ there exist two polynomials $P_n,Q_n\in\mathbb Q_p[z]$ with degree $\le n$ such that $$P_n(z)f(z)+Q_n(z)=\sum_{k\ge 2n+1}\beta_{k,n}\binom{z}{k}\quad(\beta_{k,n}\in\mathbb Q_p)$$ Is this true (known to be true??) If yes, thanks in advance for any reference or hint for the demonstration.

$\endgroup$

1 Answer 1

2
$\begingroup$

This is basically linear algebra (I assume you want $P_n$ and $Q_n$ not both zero or equivalently $P_n$ not zero) with a minor complication coming from expressing the product of two $\binom{z}{k}$'s as a sum of $\binom{z}{k}$'s.

By expressing $\binom{z}{k}$ in terms of $1$, $\frac{z-j}{j+1}$, $\frac{(z-j)(z-j-1)}{(j+1)(j+2)}$, $\dots$, $\prod_{l=j}^{j+k-1}{\frac{z-l}{l+1}}$, which is a basis of the vector space of all polynomials of degree $\leq k$, we see that $\binom{z}{k}\binom{z}{j}=\sum_{l=j}^{j+k}{a_{j,k,l}\binom{z}{l}}$ for certain rational numbers $a_{j,k,l}$.

Write $P_n(z)=\sum_{k=0}^n{p_k\binom{z}{k}}$, $Q_n(z)=\sum_{k=0}^n{q_k\binom{z}{k}}$. Then

$$P_n(z)f(z)+Q_n(z)=\sum_{l=0}^{n}{\left(\sum_{k=0}^{n}\left(\sum_{j=\max\{0,l-k\}}^{l}{\gamma_j a_{j,k,l}}\right)p_k+q_l\right)\binom{z}{l}}+\sum_{l>n}{\left(\sum_{k=0}^{n}\left(\sum_{j=\max\{0,l-k\}}^{l}{\gamma_j a_{j,k,l}}\right)p_k\right)\binom{z}{l}}.$$

Equating the first $2n+1$ coefficients on the right-hand side to zero now gives $2n+1$ linear equations in $2(n+1)$ unknowns $p_k$, $q_k$ ($k=0,\dots,n$), which have therefore a non-trivial solution.

$\endgroup$

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

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