International Journal of Statistika and Mathematika, ISSN: 2277- 2790 E-ISSN: 2249-8605, Volume 9, Issue 1, 2014 pp 11-13

Finite Difference Method for Laplace Equation

M. L. Dhumal1, S. B. Kiwne2
Department of Mathematics, Deogiri College, Station Road, Aurangabad- 431005, Maharashtra, INDIA.
Corresponding Addresses:
1 2
mldhumal19@gmail.com sbkiwne@yahoo.co.in

Research Article
Abstract: In this paper solution of Laplace equation with Dirichlet matrix equation or solution can be obtained by solving
boundary and Neumann boundary is discussed by Finite difference simultaneous equations iteratively. In this paper we use
method.
the iterative technique.
Keywords: Laplace Equation, Dirichlet boundary, Neumann
boundary, Finite Difference Approximation. 2. Finite Difference Method
2010 MSC: 65N06, 65M06, 65L12 Let us divide a two dimensional region into the points
1. Introduction with increments in the x and y direction as ∆x and ∆y
Laplace equation is a second order partial differential where ∆x=∆y=h
equation (PDE) that appears in many areas of science and Each nodal point is designated by a numbering scheme i
engineering, such as electricity, fluid flow, and steady and j where i and j are x and y increments respectively as
heat conduction. Solution of this equation, in a domain, shown in fig.2.1. We have centre difference
requires the specification of certain conditions that the approximation for second order derivatives as
unknown function must satisfy at the boundary of the (2.1)
domain. When the function itself is specified on a part of
the boundary, we call that part the Dirichlet boundary; (2.2)
when the normal derivative of the function is specified on
a part of the boundary, we call that part the Neumann By combining (2.1) and (2.2), we have
boundary. In a problem, the entire boundary can be
Dirichlet or a part of the boundary can be Dirichlet and
the rest Neumann. Consider Laplace equation in a domain Therefore
D as (2.3)

This shows that value of u at any point of the domain D is

the average of the surrounding points in the five point
With boundary condition u =f on SD and on SN, stencil of the following fig.2.1
where n is the normal to the boundary, SD is the Dirichlet
boundary, and SN is the Neumann boundary.
A typical problem is schematically shown in Fig1.1

Figure 2.1
3. Dirichlet Problem
Consider the Fig.3.1 containing only four interior
Figure 1.1 points, u is given on the East, West, North and South
Finite difference methods for partial differential walls.
equations are studied in [1],[2],[3],[4],[5],[6]. Idea of
finite difference method is to descretize the partial
differential equation by replacing partial derivatives with
their approximation that is finite differences. In this
method, the PDE is converted into a set of linear,
simultaneous equations. Which are written in the matrix
equation and then solution is obtained by solving the Figure 3.1

Copyright © 2014, Statperson Publications, Iinternational Journal of Statistika and Mathematika, ISSN: 2277- 2790 E-ISSN: 2249-8605, Volume 9 Issue 1 2014
 M. L. Dhumal, S. B. Kiwne

Thus the boundary points u(1,2), u(1,3),u(2,4), u(3,4), The Dirichlet condition is specified on the north, east and
u(4,3), u(4,2), u(3,1) and u(2,1) are known. We have to south walls. On the west wall Neumann condition is
calculate the interior points u(2,2), u(3,2), u(2,3), u(3,3). specified as:
We begin the iterative process by assuming
u(0)(2,2)= u(0)(3,2)= u(0)(2,3)= u(0)(3,3)=0
The superscript (0) denote the 0th iteration.
Using Gauss-Seidel iteration we obtain the 1st iteration We have
for interior points starting at bottom left as

Therefore
(4.1)
(3.1)
where (0,j) is the fictitious grid point out side the domain
of the problem.
We compute the first iterated values . And use it Laplace equation (2.3)at the point (1,j) becomes
in the calculation of the first iteration of , and so (4.2)

on. From (4.1) , we have

The algorithm for the iterative solution at the interior u(0,j)=u(2,j)+2hg(1,j) (4.3)
points on N x M grid Substituting value of u(0.j) from(4.3) in(4.2) we get
For i=2,N-1 -
For j=2, M-1 (4.4)
(3.2) We use equation (4.4) for 2≤j≤M-1 where g(1,j) is a
specified function.
End Do
Thus boundary values except corner points & west
End Do
boundary are known from Dirichlet boundary condition
The corner values u(1,1),u(N,1),u(1,M),u(N,M) are specified on north, east and south.
computed from The algorithm for the iterative solution is:
For i=1, N-1
(3.3)
For j=2,M-1
If i=1 then

(4.5)
Else
Thus
• The boundary values (except corner values) are
known from boundary conditions End Do
• Corner values are obtained from equation (3.3) End Do
The values of u at the corner points are computed by
• All other values are iteratively computed
using eqn.(3.3)
from(3.1) or (3.2)
Thus
4. Neumann Problem • The values u[(i,M), 2 ≤ i ≤ N-1], u[N,j), 2 ≤ j ≤
The Neumann problem is posed on the grid of fig.4.1
M-1], u[(i,1), 2 ≤ i ≤ N-1] are known from
boundary condition.
• Corner values are obtained from eqn (3.3).
• All other values are itratively computed from
eqn.(4.5)
Example 1
A Dirichlet problem is posed on a square domain as
follows:
uNorth = 100, uEast = 50, uSouth = 0, uWest = 75
Figure 4.1

International Journal of Statistiika and Mathematika, ISSN: 2277- 2790 E-ISSN: 2249-8605, Volume 9 Issue 1 Page 12
 International Journal of Statistika and Mathematika, ISSN: 2277- 2790 E-ISSN: 2249-8605, Volume 9, Issue 1, 2014 pp 11-13

When the problem is solved on a 4 x 4 grid, by using

Gauss-Seidel Iterative scheme, eight iterations are
When the problem is solved on a 4 x 4 grid, by using
necessary to reach a solution. Progress of the iterative scheme
is shown in Table 1. Gauss-Seidel Iterative scheme. The first iteration values
are shown in Table2:
Table 1: Table 2:
87.5 100 100 75.00 60 100.0 100.0 75
48.43 53.90 30 42.65 40.62 50
66.02 62.69 20 18.78 12.5 50
70.41 64.89 10 0.00 0.00 25
75.00 71.50 65.43 50.00
71.84 65.60 References
71.62 65.22 1. Smith G. D. ; Numerical solution of Partial Differential
71.42 65.39 Equations (2nd Edition), Oxford, London, 1978
71.76 65.56 2. Ames W. F. ; Numerical Methods for Partial Differential
18.75 17.19 Equations (3rd Edition), Academic Press, San Diego,
35.15 34.77 1992
43.94 39.16 3. Lapidus L and Pinder G.F. ; Numerical solution of partial
75.00 46.14 40.25 50.00 differential equations in Science and Engineering, Wiley ,
46.68 40.53 New York, 1982
46.84 40.61 4. Duchateau P. and Zachmann D. ; Applied partial
46.47 40.17 differential equations, Herper and Row, New York,1989
46.64 40.12 5. Greenspan D. and Parter S.V. ; Mildly Nonlinear Elliptic
37.50 0.00 0.00 25.00 Partial Differential Equations and their Numerical
Example 2 solution-II, Numerische Mathematik, 7, 129-146, 1965
A Neumann Problem is posed on square domain as 6. Pao C.V. ; Monotone Iterative Method fo Finite
Difference system of reaction diffusion equations,
follows Numer.math.46, 571-586, 1985.
uNorth = 100, uEast = 50, uSouth = 0, and on uwest as

Copyright © 2014, Statperson Publications, Iinternational Journal of Statistika and Mathematika, ISSN: 2277- 2790 E-ISSN: 2249-8605, Volume 9 Issue 1 2014