Derivation of Lagrange Family of Interpolation functions

Join the discussion
Registration is free. Start your own thread to ask a follow-up.
9 replies · 3K views
bugatti79
Messages
786
Reaction score
4
Folks,

I am puzzled how the linear interpolation functions (see attached) were determined based on the following equation below

##\displaystyle \psi_i=\frac{(\xi-\xi_1)(\xi-\xi_2)...(\xi-\xi_{i-1})(\xi-\xi_{i+1})...(\xi-\xi_n)}{(\xi_i-\xi_1)(\xi_i-\xi_2)...(\xi_i-\xi_{i-1})(\xi_i-\xi_{i+1})...(\xi_i-\xi_n)}##

What do the dots represent above?

and

[itex]\psi_i(\xi_j)= 1[/itex] if ##i=j## and ##0## if ##i\ne j##

For example how is ##\psi_1(\xi)## determined?

thanks
 
Attachments
  • IMAG0178.jpg
    IMAG0178.jpg
    16 KB · Views: 480
Physics news on Phys.org
bugatti79 said:
Folks,

I am puzzled how the linear interpolation functions (see attached) were determined based on the following equation below

##\displaystyle \psi_i=\frac{(\xi-\xi_1)(\xi-\xi_2)...(\xi-\xi_{i-1})(\xi-\xi_{i+1})...(\xi-\xi_n)}{(\xi_i-\xi_1)(\xi_i-\xi_2)...(\xi_i-\xi_{i-1})(\xi_i-\xi_{i+1})...(\xi_i-\xi_n)}##

What do the dots represent above?

and

[itex]\psi_i(\xi_j)= 1[/itex] if ##i=j## and ##0## if ##i\ne j##

For example how is ##\psi_1(\xi)## determined?


thanks

The dots mean fill in all the like terms for indices between the given. In this case, 3 to i-2 and i+2 to n-1.

For the second question, when j = i, the numerator and denominator are the same. When j ≠ i, one term of the numerator product = 0.
 
mathman said:
The dots mean fill in all the like terms for indices between the given. In this case, 3 to i-2 and i+2 to n-1.

For the second question, when j = i, the numerator and denominator are the same. When j ≠ i, one term of the numerator product = 0.

Sorry, I still don't follow...

I don't see how ##\psi_1(\xi)=\frac{1}{2}(1-\xi)## is obtained...?
 
bugatti79 said:
Folks,

I am puzzled how the linear interpolation functions (see attached) were determined based on the following equation below

##\displaystyle \psi_i=\frac{(\xi-\xi_1)(\xi-\xi_2)...(\xi-\xi_{i-1})(\xi-\xi_{i+1})...(\xi-\xi_n)}{(\xi_i-\xi_1)(\xi_i-\xi_2)...(\xi_i-\xi_{i-1})(\xi_i-\xi_{i+1})...(\xi_i-\xi_n)}##

What do the dots represent above?
The dots (...) are called an ellipsis, and appear as three periods. The ellipsis means "continuing in the same fashion." The first missing factor in both the numerator and denominator would be (x - x3) and the next would be (x - x4), and so on. (I don't see any purpose in writing ##\xi## when plain old x will do just fine.)
 
bugatti79 said:
I am puzzled how the linear interpolation functions (see attached) were determined based on the following equation below

##\displaystyle \psi_i=\frac{(\xi-\xi_1)(\xi-\xi_2)...(\xi-\xi_{i-1})(\xi-\xi_{i+1})...(\xi-\xi_n)}{(\xi_i-\xi_1)(\xi_i-\xi_2)...(\xi_i-\xi_{i-1})(\xi_i-\xi_{i+1})...(\xi_i-\xi_n)}##

What do the dots represent above?

like any other dots like that which you see in a mathematical expression. it means "fill in the blank" with the continuing pattern.

it's not as good as my Applied Engineering Mathematics book from Kreyszig, but the wikipedia article on Lagrange polynomials should have the answers to your question.
 
bugatti79 said:
Folks,



##\displaystyle \psi_i=\frac{(\xi-\xi_1)(\xi-\xi_2)...(\xi-\xi_{i-1})(\xi-\xi_{i+1})...(\xi-\xi_n)}{(\xi_i-\xi_1)(\xi_i-\xi_2)...(\xi_i-\xi_{i-1})(\xi_i-\xi_{i+1})...(\xi_i-\xi_n)}##



##\psi_1(\xi)=\frac{1}{2}(1-\xi)## is obtained...?

##\displaystyle \psi_1=\frac{(\xi-\xi_1)(\xi-\xi_2)(\xi-\xi_{0})(\xi-\xi_{2})(\xi-\xi_1)}{(\xi_1-\xi_1)(\xi_1-\xi_2)(\xi_1-\xi_{0})(\xi_1-\xi_{2})(\xi_1-\xi_1)}##...?
 
bugatti79 said:
##\displaystyle \psi_1=\frac{(\xi-\xi_1)(\xi-\xi_2)(\xi-\xi_{0})(\xi-\xi_{2})(\xi-\xi_1)}{(\xi_1-\xi_1)(\xi_1-\xi_2)(\xi_1-\xi_{0})(\xi_1-\xi_{2})(\xi_1-\xi_1)}##...?

ξ1 - ξ1 is NOT a term in the denominator. Look at the definition of ψi, the ith term is left out of the numerator and denominator.

Where did you get the expression for ψ1(ξ)?
 
mathman said:
ξ1 - ξ1 is NOT a term in the denominator. Look at the definition of ψi, the ith term is left out of the numerator and denominator.

Where did you get the expression for ψ1(ξ)?

Its the first equation shown in the picture in first post...
 
bugatti79 said:
Its the first equation shown in the picture in first post...

The pictures are slightly criptic. The important thing is that the definition of ψi depends on n. The particular item you ask about is for n = 2.
In that case:
ψ1(ξ) = (ξ-ξ2)/(ξ1-ξ2), where ξ1= -1 and ξ2=+1.
 
mathman said:
The pictures are slightly criptic. The important thing is that the definition of ψi depends on n. The particular item you ask about is for n = 2.
In that case:
ψ1(ξ) = (ξ-ξ2)/(ξ1-ξ2), where ξ1= -1 and ξ2=+1.

OK, thanks