Bounded functions with unbounded integrals

Join the discussion
Registration is free. Ask a follow-up in this thread, or start your own.
3 replies · 3K views
roam
Messages
1,265
Reaction score
12

Homework Statement



I am trying to show that the integrator is unstable by giving examples of bounded inputs which produce unbounded outputs (i.e. a bounded function whose integral is unbounded).

Note: The integrator is a system which gives an output equal to the anti-derivative of its input.

The Attempt at a Solution



I have already proven the instability of the differentiator by considering the bounded input ##f(t)= \sin(t^2),## which gives the unbounded output ##f'(t) = 2t \cos(t^2).##

23h8yzp.jpg


For the integrator, I know, for instance, that the bounded input ##f(t)=1## gives the unbounded output ##t.## But could anyone suggest a more interesting example like the one I gave for the differentiator?

I couldn't come up with a good example. I would appreciate any suggestions or links.
 
Physics news on Phys.org
roam said:

Homework Statement



I am trying to show that the integrator is unstable by giving examples of bounded inputs which produce unbounded outputs (i.e. a bounded function whose integral is unbounded).

Note: The integrator is a system which gives an output equal to the anti-derivative of its input.

The Attempt at a Solution



I have already proven the instability of the differentiator by considering the bounded input ##f(t)= \sin(t^2),## which gives the unbounded output ##f'(t) = 2t \cos(t^2).##

View attachment 139989

For the integrator, I know, for instance, that the bounded input ##f(t)=1## gives the unbounded output ##t.## But could anyone suggest a more interesting example like the one I gave for the differentiator?

I couldn't come up with a good example. I would appreciate any suggestions or links.
Your problem might be more interesting on a bounded interval. Think about ##\sqrt[3] x## on ##(0,1)## for a differentiation example. But on a bounded interval no example exists for the integration because ##\left | \int_a^t f(u)~du\right | \le \int_a^b |f(u)|~du\le M(b-a)## where ##M## is a bound for ##|f|## on ##[a,b]##.
 
LCKurtz said:
Your problem might be more interesting on a bounded interval. Think about ##\sqrt[3] x## on ##(0,1)## for a differentiation example. But on a bounded interval no example exists for the integration because ##\left | \int_a^t f(u)~du\right | \le \int_a^b |f(u)|~du\le M(b-a)## where ##M## is a bound for ##|f|## on ##[a,b]##.

What about a non-bounded interval? I mean, I already found a function, ##f(t)=constant##, which increases indefinitely. The only other example I can think of, would be the Heaviside step function, ##f(t)=u(t)##. The output of the integrator for this function is:

$$g(t) = \intop^t_{-\infty} u(\tau) d\tau = \intop^t_{-\infty} 1 \ d\tau = t$$

for ##t>0.##

So, is there really no other function whose integral increases indefinitely? :confused:
 
roam said:
What about a non-bounded interval? I mean, I already found a function, ##f(t)=constant##, which increases indefinitely. The only other example I can think of, would be the Heaviside step function, ##f(t)=u(t)##. The output of the integrator for this function is:

$$g(t) = \intop^t_{-\infty} u(\tau) d\tau = \intop^t_{-\infty} 1 \ d\tau = t$$

for ##t>0.##

So, is there really no other function whose integral increases indefinitely? :confused:
Of course not. There are lots of examples. Take any function that is nonnegative for ##x\ge 0## that is bounded but has an infinite area. One such example is ##\arctan x,~0\le x## which is bounded by ##\frac \pi 2##.
 
  • Like
Likes   Reactions: roam