Home Page Overview Site Map Index Appendix Illustration About Contact Update FAQ

## Mathematical Minimum (for Aspiring Physicists)

### Numerical Solution for 2nd Order Differential Equations

In general, a 2nd order differential equation involves the independent variable x, the dependent variable y, its 1st derivative dy/dx, and the 2nd derivative d(dy/dx)/dx. Numerical solution requires only simple programming whenever we can isolate the 2nd derivative in the form:

d(dy/dx)/dx=f(y',y,x),
where y'=dy/dx=z.
Thus, we can rewrite the original equation as:
dz/dx=f(z,y,x).

In the form of difference equation:
z(n+1)=z(n)+f[z(n),y(n),x(n)]dx
y(n+1)=y(n)+z(n)dx
Input parameters include the step size dx, the number of steps max, and and the initial conditions x(0), y(0), z(0).

Numerical integration is carried out by iteration as shown below (in Basic, also see actual example for a cosmological model):

input max
input dx
input x0
input y0
input z0
n=0
zn=z0
yn=y0
xn=x0
#from the actual expression of the function at x0
fn=f0
MainLoop:
znn=zn+fn*dx
ynn=yn+zn*dx
xnn=xn+dx
n=n+1
if (n-max)=0 then goto Ende
zn=znn
yn=ynn
xn=xnn
#from the actual expression of the function
fn=fnn
print xnn;:print ", ";:print ynn
refresh
goto MainLoop
Ende:

The numerical solution for 1st order differential equation follows the same idea only simpler (see a Basic Program for the Logistic Equation). Other "home-made" programs such as Base Conversion (between decimal, hexadecimal, and binary), Unit Conversion (fps to SI, SI to fps), Mathematical Functions, and Loan Payment (all in Basic) are available (just clicking the link) to show the subtlety of computer programming.

.