I have $a_1, ..., a_n, b_1, ..., b_n \in \mathbb{R}$. I must show that there is a unique polynomial $f(x) = c_0 + c_1x +...+ c_{n-1}x^{n-1} \in \mathbb{R}[x]$ of degree less than $n$ such that $f(a_1) = b_1, ..., f(a_n) = b_n$

$a_i$ is distinct and $b_i$ are just any real numbers.

Now, I am not sure how to quite do this, but I think that the proof might be easy if you use the Vandermonde determinant \begin{align*} A= \begin{pmatrix} 1 & x_1 & x_1^2 & \cdots & x_1^{n-1} \\ 1 & x_2 & x_2^2 & \cdots & x_2^{n-1} \\ \vdots \\ 1 & x_n & x_n^2 & \cdots & x_n^{n-1} \end{pmatrix} \end{align*} \begin{align*} \det(A) = \prod_{j > i} (x_j - x_i). \end{align*}

However, I am not sure how to carry out this proof. Is my thought process correct? Could somebody tell me how should I go on from here? Thanks.

  • $\begingroup$ Why do you think the proof might be easy using that determinant? Did someone give you a hint? $\endgroup$ – darij grinberg Nov 14 '18 at 3:18
  • $\begingroup$ I just think that multiplying that matrix to a vector of variables would give the linear system that shows $f(a_1)=b_1, ..., f(a_n)=b_n$ $\endgroup$ – dmsj djsl Nov 14 '18 at 3:19
  • 1
    $\begingroup$ Good, but what vector? And what should the $x_i$ be in the matrix? $\endgroup$ – darij grinberg Nov 14 '18 at 3:22
  • $\begingroup$ I was thinking about keeping the $x$s in the matrix and multiplying by the vector of $c_i$. Is this not okay? If this is correct, I am still stuck on how to prove that this polynomial is unique. Am I missing something obvious? $\endgroup$ – dmsj djsl Nov 14 '18 at 3:48
  • $\begingroup$ No. What does keeping the $x$s mean? You don't have any $x$'s given. You want $b_i = f\left(a_i\right) = c_0 + c_1 a_i^1 + \cdots + c_{n-1} a_i^{n-1}$ for all $i$; does this suggest anything about what matrix and what vectors you should take? $\endgroup$ – darij grinberg Nov 14 '18 at 3:49

Consider the map $\phi:\def\R{\Bbb R}\R[x]_{<n}\to\R^n$ defined by $\phi(P)=(P[a_1],P[a_2],\ldots,P[a_n])$, where $\R[x]_{<n}$ is the subspace of $\R[x]$ of polynomials of degree less than$~n$, and $P[a]$ denotes the evaluation of the polynomial $P$ at $x=a$. Your are asked to show that $\phi$ is a bijection: every point $(b_1,\ldots,b_n)$ equals $\phi(P)$ for a unique $P$.

Now $\phi$ is a linear map, since every evaluation map $P\mapsto P[a_i]$ is linear. Also, clearly $\phi(x^k)=(a_1^k,a_2^k,\ldots,a_n^k)$ for $k=0,1,\ldots,n-1$, so the matrix $M$ of $\phi$ with respect to the basis $[1,x,x^2,\ldots,x^{n-1}]$ of $\R[x]_{<n}$ is essentially just matrix $A$ of the question, where one just needs to set $x_i:=a_i$ for $i=1,2,\ldots,n$. The evaluation of $\det(A)$ that you cited, and the fact that the $a_i$ are all distinct (i.e., $a_j-a_i\neq0$ whenever $i<j$) show that $\det(M)\neq0$. Then $M$ is an invertible matrix, which corresponds to the fact that $\phi$ is a bijective linear map, and you are done.


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.