Take the 2-minute tour ×
Mathematics Stack Exchange is a question and answer site for people studying math at any level and professionals in related fields. It's 100% free, no registration required.

I hesitate here because of an understanding with a calculation problems.

I want to calculate an interpolation using the Vandermonde matrix. see: http://en.wikipedia.org/wiki/Vandermonde_matrix My polynomial has degree 2. I have the following coordinates p1 (1,4), p2 (2,6), p3 (4,1). I want a formula exhibition that interpolates these points. An interpolation point is difend as $p(u,g_i)$

The general matrix form is like this: T*M*vector_g

$q(u)= \begin{matrix} (u ^ {2} & u & 1) \end{matrix} * \begin{matrix} a_1 & a_2 &a_3 \\b_2 & b_2 &b_2 \\c_1 & c_1 &c_2 \end{matrix} * \begin{matrix} g_0\\g_1\\g_2 \end{matrix} $

Explenation:

Monomvector T = ($u^{2}$ u 1 )

geometry vector vector_g = $\begin{matrix} g_0\\g_1\\g_2 \end{matrix} $

Vanemonde Matrix M = $\begin{matrix} a_1 & a_2 &a_3 \\b_2 & b_2 &b_2 \\c_1 & c_1 &c_2 \end{matrix} $

Now my understanding problem:

When i use the given parameter p1..p2 i have a linear equation like

p1 (1,4) => a + b + c = 4 p2 (2,6) => 4a + 2b + c = 6 p3 (4,1) => 16a + 4b + c = 4

and the result are the coefficients for a,b,c are $\frac{-3}{2} u^2 $\frac{13}{2}u -1$

But my Professor wants the general matrix form like above .... all i have is following :

q(u) = ($u^{2}$ u 1) * $(\frac{-3}{2} u^2 $\frac{13}{2}u -1)$ = g0

But when i calculate

p1 (1,4) => a + b + c = $g_0$

p2 (2,6) => 4a + 2b + c = $g_1$

p3 (4,1) => 16a + 4b + c = $g_2$

i get a wrong matrix but in the right form ....

if you need more information just tell

share|improve this question
1  
Although I believe I understand the problem and will try to write an answer, I ultimately find myself unable to figure out a reasonable way to edit the original question. In the future please consider using grammar, spelling, capitalization, and more readable math so that more than just a handful of people can understand you. –  anon Aug 4 '11 at 15:02
    
some paragraphs are gone... –  RenHoek Aug 4 '11 at 15:23

1 Answer 1

up vote 1 down vote accepted

The purpose of polynomial interpolation is to find a polynomial $p(x)$ whose graph touches a given set of points and therefore gives a reasonable curve to approximate data. In this case, we want a polynomial with the following conditions:

$$p(1)=4 , \quad p(2)=6, \quad p(4)=1.$$

This comes from the fact that if $(x,y)$ is a point on the graph, $p(x)=y$ is a condition on the polynomial. The number of points is $3$, so the degree of our polynomial is at most $2$ (quadratic). If we let $p(x)=fx^2+gx+h$, then (plugging this into the three conditions) this gives the system

$$f+g+h=4$$ $$4f+2g+h=6$$ $$ 16f+4g+h=1.$$

Putting this in matrix form:

$$\begin{pmatrix} 1 & 1 & 1\\ 4 & 2 & 1\\ 16 & 4 & 1\end{pmatrix} \begin{pmatrix}f\\ g\\ h\end{pmatrix} = \begin{pmatrix}4\\ 6\\ 1\end{pmatrix}.$$

Hence

$$ p(x) = \begin{pmatrix}x^2 & x & 1\end{pmatrix}\begin{pmatrix}f\\g\\h\end{pmatrix}=\begin{pmatrix}x^2 & x & 1\end{pmatrix}\begin{pmatrix} 1 & 1 & 1\\ 4 & 2 & 1\\ 16 & 4 & 1\end{pmatrix}^{-1}\begin{pmatrix}4\\ 6\\ 1\end{pmatrix}$$ $$=\begin{pmatrix}x^2 & x & 1\end{pmatrix}\begin{pmatrix} 1/3 & -1/2 & 1/6\\ -2 & 5/2 & -1/2\\ 8/3 & -2 & 1/3\end{pmatrix}\begin{pmatrix}4\\ 6\\ 1\end{pmatrix}= \begin{pmatrix}x^2 & x & 1\end{pmatrix}\begin{pmatrix}-3/2\\13/2\\-1\end{pmatrix}$$

$$ = -\frac{3}{2}x^2+\frac{13}{2}x-1.$$

You can check that this is the correct polynomial. If you have any questions about the matrix manipulations, such as how I calculated the inverse, I'll see what I can do.

share|improve this answer
    
why are calculation here $p(x) = \begin{pmatrix}x^2 & x & 1\end{pmatrix}\begin{pmatrix}f\\g\\h\end{pmatrix}=\begin{pmatrix}x^2 & x & 1\end{pmatrix}\begin{pmatrix} 1 & 1 & 1\\ 4 & 2 & 1\\ 16 & 4 & 1\end{pmatrix}^{-1}\begin{pmatrix}4\\ 6\\ 1\end{pmatrix}$ the inverse matrix ??? –  RenHoek Aug 4 '11 at 15:33
1  
@RenHoek: If you have $Af=b$, in order to find $f$, you need to use the inverse with $f=A^{-1}b$. This holds true with matrices as it does with scalars. –  anon Aug 4 '11 at 15:38
    
okay,.. it's and matrix "division" like $x^{-1} = \frac{1}{x}$ –  RenHoek Aug 4 '11 at 15:40
    
@RenHoek: Yes. Do you know what a matrix inverse is? I suggest you familiarize yourself with basic linear algebra more if your class is going to use it this much consistently. –  anon Aug 4 '11 at 15:43
    
i calculated a lot to test your result way... it complete the right way :) thanks a lot ! you made my day –  RenHoek Aug 4 '11 at 16:19

Your Answer

 
discard

By posting your answer, you agree to the privacy policy and terms of service.

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