Find Rate of Convergence for sin(h)/h

  • Thread starter Thread starter stunner5000pt
  • Start date Start date
  • Tags Tags
    Convergence
Join the discussion
Ask a follow-up here, or get your own question answered by working scientists, mathematicians and engineers — people, not an autocomplete.
Real named experts · corrections over time · the nuance an AI answer skips
8 replies · 3K views
stunner5000pt
Messages
1,447
Reaction score
5
Find the rate of convergence of
[tex]\lim_{h \rightarrow 0}(\frac{\sin(h)}{h}) = 1[/tex]

well I am not really sure on what to do
[tex]\sin(h) \leq 1[/tex]
[tex]\frac{\sin(h)}{h} \leq \frac{1}{h}[/tex]
so then sine converges with a rate of O(1/h) ?
but the answer in the book is O(h^2) how so?


please help! Thank you
 
on Phys.org
hi,

I'm doing some of the same stuff in my analysis class, and in my class notes the teacher wrote that sin(x) = O(x), not x^2.

Going back to the definitions of 'little o' and 'big O' might help.

f = O(g) means that the ratio of f/g is bounded by some constant, where f = o(g) means that f aproaches zero faster than g.

Sorry that's not a complete answer, but I'm still learning this stuff too.

Hope it helps :blushing:
 
You managed to show sin(h)/h=O(1/h) as h->0, but this isn't a very useful bound since 1/h is diverging. To get the rate of convergence of that limit, it's not sin(h)/h you are trying to bound, but

[tex]\left|\frac{\sin h}{h}-1\right|[/tex]
 
shmoe said:
You managed to show sin(h)/h=O(1/h) as h->0, but this isn't a very useful bound since 1/h is diverging. To get the rate of convergence of that limit, it's not sin(h)/h you are trying to bound, but
[tex]\left|\frac{\sin h}{h}-1\right|[/tex]

ok i see what you mean ... that the stuff of the right must converge to same thing as the sequence does

now for playing around with [tex]\left|\frac{\sin h}{h}-1\right|[/tex]
or is there a systematic method here?

[tex][tex]\left|\frac{\sin h}{h}-1\right| \leq \frac{1}{h} -1 = \frac{1-h}{h} =[/tex] not sure what to do here though...[/tex]
 
Don't get rid of the sin(h), it's a key ingredient here. Again, it's no use bounding this thing by something that diverges as h->0, though I'd question how exactly you removed the absolute value sign.

How to proceed depends on what you're allowed to use, it's pretty trivial with taylor series say, or by looking at the limits with l'hospital's rule.
 
how would i use L'hopital's rule to find the rate of convergence here??

when using taylor series would i use Xo = 0 ?? i.e. Maclaurin series?

in which case i do get the O(h^2) term for the first taylor polynomial
so in other words for questions like this should i lawyas use the first taylor polynomial? What about the second or third ..?