Computational Physics - Department of Physics

(Axel Boer) #1

208 6 Linear Algebra


derive the following equation for potentialV

d^2 V
dx^2

=

V√^3 /^2

x

,

withV( 0 ) = 1.
In our case we will rewrite Poisson’s equation in terms of dimensionless variables. We can
then rewrite the equation as

−u′′(x) =f(x), x∈( 0 , 1 ), u( 0 ) =u( 1 ) = 0.

and we define the discretized approximation touasviwith grid pointsxi=ihin the interval
fromx 0 = 0 toxn+ 1 = 1. The step length or spacing is defined ash= 1 /(n+ 1 ). We have then
the boundary conditionsv 0 =vn+ 1 = 0. We approximate the second derivative ofuwith


vi+ 1 +vi− 1 − 2 vi
h^2
=fi fori= 1 ,...,n,

wherefi=f(xi). Show that you can rewrite this equation as a linear set of equations of the
form
Av=b ̃,
whereAis ann×ntridiagonal matrix which we rewrite as

A=









2 −1 0 ... ... 0

−1 2 −1 0 ... ...

0 −1 2 −1 0 ...

... ... ... ... ...

0 ... −1 2 − 1

0 ... 0 −1 2









andb ̃i=h^2 fi.
In our case we will assume thatf(x) = ( 3 x+x^2 )ex, and keep the same interval and boundary
conditions. Then the above differential equation has an analytic solution given byu(x) =
x( 1 −x)ex(convince yourself that this is correct by inserting the solution in the Poisson
equation). We will compare our numerical solution with thisanalytic result in the next
exercise.


  1. We can rewrite our matrixAin terms of one-dimensional vectorsa,b,cof length1 :n. Our
    linear equation reads


A=









b 1 c 1 0 ... ... ...
a 2 b 2 c 2 ... ... ...
a 3 b 3 c 3 ... ...
... ... ... ... ...
an− 2 bn− 1 cn− 1
an bn

















v 1
v 2
...
...
...
vn









=









b ̃ 1
b ̃ 2
...
...
...
b ̃n









.

A tridiagonal matrix is a special form of banded matrix whereall the elements are zero
except for those on and immediately above and below the leading diagonal. The above
tridiagonal system can be written as

aivi− 1 +bivi+civi+ 1 =b ̃i,
Free download pdf