1.7: Applications to Laplace transforms
( \newcommand{\kernel}{\mathrm{null}\,}\)
The Laplace transform of a function f locally integrable on [0,∞) is
F(s)=∫∞0e−sxf(x)dx
for all s such that integral converges. Laplace transforms are widely applied in mathematics, particularly in solving differential equations.
We leave it to you to prove the following theorem (Exercise [exer:26]).
[theorem:12] Suppose f is locally integrable on [0,∞) and |f(x)|≤Mes0x for sufficiently large x. Then the Laplace transform of F converges uniformly on [s1,∞) if s1>s0.
[theorem:13] If f is continuous on [0,∞) and H(x)=∫∞0e−s0uf(u)du is bounded on [0,∞), then the Laplace transform of f converges uniformly on [s1,∞) if s1>s0.
If 0≤r≤r1,
∫r1re−sxf(x)dx=∫r1re−(s−s0)xe−s0xf(x)dt=∫r1re−(s−s0)tH′(x)dt.
Integration by parts yields
∫r1re−sxf(x)dt=e−(s−s0)xH(x)|r1r+(s−s0)∫r1re−(s−s0)xH(x)dx.
Therefore, if |H(x)|≤M, then
|∫r1re−sxf(x)dx|≤M|e−(s−s0)r1+e−(s−s0)r+(s−s0)∫r1re−(s−s0)xdx|≤3Me−(s−s0)r≤3Me−(s1−s0)r,s≥s1.
Now Theorem [theorem:4] implies that F(s) converges uniformly on [s1,∞).
The following theorem draws a considerably stonger conclusion from the same assumptions.
[theorem:14] If f is continuous on [0,∞) and
H(x)=∫x0e−s0uf(u)du
is bounded on [0,∞), then the Laplace transform of f is infinitely differentiable on (s0,∞), with
F(n)(s)=(−1)n∫∞0e−sxxnf(x)dx;
that is, the n-th derivative of the Laplace transform of f(x) is the Laplace transform of (−1)nxnf(x).
First we will show that the integrals
In(s)=∫∞0e−sxxnf(x)dx,n=0,1,2,…
all converge uniformly on [s1,∞) if s1>s0. If 0<r<r1, then
∫r1re−sxxnf(x)dx=∫r1re−(s−s0)xe−s0xxnf(x)dx=∫r1re−(s−s0)xxnH′(x)dx.
Integrating by parts yields
∫r1re−sxxnf(x)dx=rn1e−(s−s0)r1H(r)−rne−(s−s0)rH(r)−∫r1rH(x)(e−(s−s0)xxn)′dx,
where ′ indicates differentiation with respect to x. Therefore, if |H(x)|≤M≤∞ on [0,∞), then
|∫r1re−sxxnf(x)dx|≤M(e−(s−s0)rrn+e−(s−s0)rrn+∫∞r|(e−(s−s0)x)xn)′|dx).
Therefore, since e−(s−s0)rrn decreases monotonically on (n,∞) if s>s0 (check!),
|∫r1re−sxxnf(x)dx|<3Me−(s−s0)rrn,n<r<r1,
so Theorem [theorem:4] implies that In(s) converges uniformly [s1,∞) if s1>s0. Now Theorem [theorem:11] implies that Fn+1=−F′n, and an easy induction proof yields [eq:30] (Exercise [exer:25]).
[example:13] Here we apply Theorem [theorem:12] with f(x)=cosax (a≠0) and s0=0. Since
∫x0cosaudu=sinaxa
is bounded on (0,∞), Theorem [theorem:12] implies that
F(s)=∫∞0e−sxcosaxdx
converges and
F(n)(s)=(−1)n∫∞0e−sxxncosaxdx,s>0.
(Note that this is also true if a=0.) Elementary integration yields
F(s)=ss2+a2.
Hence, from [eq:31],
∫∞0e−sxxncosax=(−1)ndndsnss2+a2,n=0,1,….