Are x, y, and z elements of the normed linear space ℓ∞(ℝ)?

  • Thread starter Thread starter bugatti79
  • Start date Start date
  • Tags Tags
    Linear Space
Join the discussion
Registration is free. Start your own thread to ask a follow-up.
32 replies · 8K views
bugatti79
Messages
786
Reaction score
4

Homework Statement



Consider the normed linear space [itex]\ell_\infty \mathbb({R})[/itex]. Let [itex]x= \frac{n-1}{n}[/itex], [itex]y=(1/n) and z=2^n[/itex]

a) Are x,y,z each in [itex]\ell_\infty \mathbb({R})[/itex]

b) What is x+y? What is [itex]2^{1/2} y?[/itex]

c) Calculate [itex]||x||_\infty[/itex], [itex]||y||_\infty[/itex], [itex]||x+y||_\infty[/itex] and [itex]||2^{1/2} y ||_\infty[/itex]

For a) Does that mean are x,y,z subspaces of [itex]\ell_\infty \mathbb({R})[/itex]?
 
Physics news on Phys.org
bugatti79 said:

Homework Statement



Consider the normed linear space [itex]\ell_\infty \mathbb({R})[/itex]. Let [itex]x= \frac{n-1}{n}[/itex], [itex]y=(1/n) and z=2^n[/itex]

a) Are x,y,z each in [itex]\ell_\infty \mathbb({R})[/itex]

b) What is x+y? What is [itex]2^{1/2} y?[/itex]

c) Calculate [itex]||x||_\infty[/itex], [itex]||y||_\infty[/itex], [itex]||x+y||_\infty[/itex] and [itex]||2^{1/2} y ||_\infty[/itex]

For a) Does that mean are x,y,z subspaces of [itex]\ell_\infty \mathbb({R})[/itex]?
No, x, y, and z are sequences. The question is asking whether they belong to [itex]\ell_\infty \mathbb({R})[/itex].
 
x and y are each an element in l_∞ as they are bounded above

z is NOT an element as its not bounded above

In b) what does it mean by 'what is'. Is that asking to evaluate?
 
bugatti79 said:
1. Homework Statement

b) What is x+y? What is [itex]2^{1/2} y?[/itex]



x+y =n implies (n) =(1,2,3,4...) which is not in l∞(R) sicne it is not bounded above.

2^(1/2)*y= 2^(n+1/2) which is not in l∞(R) either...?
 
bugatti79 said:
x+y =n implies (n) =(1,2,3,4...) which is not in l∞(R) sicne it is not bounded above.
Yes, but here's how to say the first part better. x + y = {n} = {1, 2, 3, ...}
bugatti79 said:
2^(1/2)*y= 2^(n+1/2) which is not in l∞(R) either...?
No, 21/2y = 21/2{1/n}. That means 21/2 times each element of the sequence {1/n}.
 
Mark44 said:
No, 21/2y = 21/2{1/n}. That means 21/2 times each element of the sequence {1/n}.

Thanks..that was a mistake on my part. So hence it IS in l∞(R)

For c) do I attempt to evaluate each?
 
Yes and yes.
Your textbook should have a definition of the infinity norm - if not, there's one in the link I posted in the other thread.
 
I like Serena said:
Hmm, I get a different result for x+y... :confused:

Mark44 said:
Yes, but here's how to say the first part better. x + y = {n} = {1, 2, 3, ...}


Let's make that x + y = {1} = {1, 1, 1, ...}
 
bugatti79 said:

Homework Statement



Consider the normed linear space [itex]\ell_\infty \mathbb({R})[/itex]. Let [itex]x= \frac{n-1}{n}[/itex], [itex]y=(1/n) and z=2^n[/itex]

c) Calculate [itex]||x||_\infty[/itex], [itex]||y||_\infty[/itex], [itex]||x+y||_\infty[/itex] and [itex]||2^{1/2} y ||_\infty[/itex]

For x:

1-1/n =(0,1/2,2/3,3/4...) and is bounded.

[itex]||1-1/n||_∞[/itex]=...? I am not sure how to calculate this. If some one can illustrate this one, I can try the others...? I know [itex]||1-1/n||_∞=sup |x_n|[/itex] but don't know how to calculate |x_n|

Thanks
 
Mark44 said:
See http://en.wikipedia.org/wiki/Lp_space, in the section titled Lp spaces, starting with "One also defines the ∞-norm as ...".

I believe that this is the norm to be used in your problem.

THanks Mark but how to I evalute |x_n|?
 
Mark44 said:
What is the supremum of the set {0, 1/2, 2/3, 3/4, ..., n/(n + 1), ...}?

I think the supremum is the smallest real number that is greater than or equal to every number in the set, hence t must be 0...?

What do i do next?

Is it something along the lines of

[itex]|1-1/n|=\sqrt{(1-1/n)^2}=0[/itex] where n=1
 
Last edited:
bugatti79 said:
I think the supremum is the smallest real number that is greater than or equal to every number in the set,
Yes.
bugatti79 said:
hence t must be 0...?
No. 0 is smaller than, not larger than every number in the sequence except the first.
bugatti79 said:
What do i do next?

Is it something along the lines of

[itex]|1-1/n|=\sqrt{(1-1/n)^2}=0[/itex] where n=1
 
Mark44 said:
Yes.No. 0 is smaller than, not larger than every number in the sequence except the first.
Then it doesn't exist, there is no sup...?
 
bugatti79 said:
I think the supremum is the smallest real number that is greater than or equal to every number in the set
Yes! So what is the smallest number that is >= each number in the sequence {0, 1/2, 2/3, 3/4, ..., n/(n+1), ...}
bugatti79 said:
, hence t must be 0...?
No, it can't be 0. This number is too small, because it's not larger than 1/2 (for example). 2589 is larger than all the numbers in the sequence, but can you find one that is not so large?
 
Mark44 said:
Yes! So what is the smallest number that is >= each number in the sequence {0, 1/2, 2/3, 3/4, ..., n/(n+1), ...}
No, it can't be 0. This number is too small, because it's not larger than 1/2 (for example). 2589 is larger than all the numbers in the sequence, but can you find one that is not so large?

1 would be greater than or equal to every number in the set...?
 
bugatti79 said:
1 would be greater than or equal to every number in the set...?

a reasonable guess, can you prove it?
 
bugatti79 said:
1 would be greater than or equal to every number in the set...?
Yes. Why are you uncertain?

Deveno said:
a reasonable guess, can you prove it?
Although not required in the problem, Deveno's question is reasonable, given the uncertainty in the answer above.
 
Mark44 said:
Yes. Why are you uncertain?

Although not required in the problem, Deveno's question is reasonable, given the uncertainty in the answer above.

All I know is that |1- 1/n| < 1 but I don't know to prove its 1.
 
|1 - 1/n| is never 1.

but...perhaps there might be a least upper bound. what are our likely candidates for this office?
 
I like Serena said:
What is:
[tex]\lim\limits_{n \to \infty} (1 - {1 \over n})[/tex]

IF n tends to infinity then the modulus tends to 1 but never 1...so does that mean that
||1- 1/n||_∞=sup |1-1/n|=1?
 
I don't understand your uncertainty either.

I'm just going to say yes, it is 1.
 
bugatti79 said:
IF n tends to infinity then the modulus tends to 1 but never 1...so does that mean that
||1- 1/n||_∞=sup |1-1/n|=1?

what does:
[tex]\lim_{n \to \infty} f(n) = L[/tex]
mean?

if you have such a limit, and, you know that f(n) < L for all n, what must sup{f(n)} be?
 
Deveno said:
what does:
[tex]\lim_{n \to \infty} f(n) = L[/tex]
mean?

if you have such a limit, and, you know that f(n) < L for all n, what must sup{f(n)} be?

That L is a finte real number and in this case it is a 1.
 
Deveno said:
what does:
[tex]\lim_{n \to \infty} f(n) = L[/tex]
mean?

if you have such a limit, and, you know that f(n) < L for all n, what must sup{f(n)} be?

bugatti79 said:
That L is a finte real number and in this case it is a 1.
You didn't answer Deveno's question. He asked, what is sup{f(n)}?
 
Mark44 said:
You didn't answer Deveno's question. He asked, what is sup{f(n)}?

Is it L?