If you are doing Laplace transforms (classic place for t and s) you get square root of Pi/s. Maple output:
> with(inttrans);
[addtable, fourier, fouriercos, fouriersin, hankel, hilbert,
invfourier, invhilbert, invlaplace, laplace, mellin]
> laplace(t^(-1/2),t,s);
1/2
Pi
-----
1/2
s
2006-12-19 11:58:57
·
answer #1
·
answered by a_math_guy 5
·
0⤊
0⤋
The indefinite integral of that function is only expressible using non-elementary functions, such as the error function. However, given the form, I suspect (as the other posters have mentioned) that you are doing a Laplace transform of 1/ât, in which case we only need the definite integral:
[0, â]â«t^(-1/2)e^(-st) dt
In this case, consider the substitution:
t=u²/s, dt=2u/s du
Which gives:
[0, â]â«1/ât e^(-u²) 2u/s du (note that this particular substitution does not change the limits of integration)
But 1/ât = âs/u, so this is:
[0, â]â«2/âs e^(-u²) du
2/âs [0, â]â«e^(-u²) du
But clearly, the function e^(-u²) is symmetric about the y-axis, so [0, â]â«e^(-u²) du = [-â, 0]â«e^(-u²) du = 1/2 [-â, â]â«e^(-u²) du. Thus you have:
1/âs [-â, â]â«e^(-u²) du
And the latter quantity is of course the gaussian integral, with value âÏ. Thus you have:
[0, â]â«t^(-1/2)e^(-st) dt = â(Ï/s)
ETA: for a proof of the value of the gaussian integral, see the wikipedia article http://en.wikipedia.org/wiki/Gaussian_integral
2006-12-19 21:22:08
·
answer #2
·
answered by Pascal 7
·
2⤊
0⤋
Are you doing a laplace transform?
When I tried just to get the general answer I get an infinite series that is not any recognizable function to me.
However if s=1 and and improper integral is taken over the positive x
domain the function is the gamma function of 1/2 which is the square root of pi = 1.77245.... If this rings a bell then look into the gamma function.
2006-12-19 20:53:08
·
answer #3
·
answered by adrian b 3
·
0⤊
0⤋
integral of udv = uv - int(v)du
u = t^-1/2
du = -1/2t^-3/2
v=e^(-st)
dv = 1/-s(e^-st)
write in the that form as above with those values evualate and your done. I hate doing math on computer
2006-12-19 20:11:10
·
answer #4
·
answered by onewhosubmits 6
·
0⤊
0⤋
well i think you hav to use integration by parts
2006-12-19 19:58:11
·
answer #5
·
answered by gamefreak 3
·
0⤊
0⤋