Proving the Convergence of a Sequence: Monotonicity and Boundedness

In summary, we have a sequence x_n=\sum_{k=1}^{k=n}{\frac{1}{2^{\sqrt{k}}}} that we need to prove is convergent. We use the integral test and the fact that as n approaches infinity, 2n is always greater than n^4. Using this, we prove that the sequence is increasing and has an upper bound, thus it is convergent. This is done by showing that for any natural n greater than a certain value, the sum of the series is less than a finite value. This proves the convergence of the sequence.
  • #1
paniurelis
12
0
Let's have a sequence [tex]x_n=\sum_{k=1}^{k=n}{\frac{1}{2^{\sqrt{k}}}[/tex].
We must prove it is convergent.
First thought, let's try to prove it is monotonic and bounded, which means convergence of sequence.
Monotonicity is easy, [tex]\forall n \in N: x_{n+1}-x_n = \frac{1}{2^{\sqrt{n+1}}} > 0[/tex]
So, sequence is increasing. Next, I should prove it has an upper bound, but I am not able to come up with bigger sequence which would have a positive finite limit.
Any ideas ?
 
Physics news on Phys.org
  • #2
I think the best way to approach this problem is via the integral test. 2^(-sqrt(k)) is going very slowly to zero, so it's going to be tough to find anything that goes to zero slower than it but for which we can verify summability easily. (But it's 5:30am, so I'm probably missing something!)
 
  • #3
Thanks morphism, it is a good idea. The problem is, I have an impression, the solution should need only the knowledge of limit theory and some elementary algebra, because I found this problem in the problem book for calculus I, before the chapters with problems about integrals...
IMHO, there is some clever trick I am not aware about...:-(
So, does anybody have more ideas ?
 
  • #4
paniurelis said:
Thanks morphism, it is a good idea. The problem is, I have an impression, the solution should need only the knowledge of limit theory and some elementary algebra, because I found this problem in the problem book for calculus I, before the chapters with problems about integrals...
IMHO, there is some clever trick I am not aware about...:-(
So, does anybody have more ideas ?

So for any natural n large enough (say, n >= N), we have: 2n >= n4. You can prove this using the Limit: [tex]\lim_{x \rightarrow \infty} \frac{x ^ 4}{2 ^ x} = 0[/tex].

So, for any natual n > N, we have:

[tex]\sum_{k = 1} ^ {n} \frac{1}{2 ^ {\sqrt{k}}} = \sum_{k = 1} ^ {N} \frac{1}{2 ^ {\sqrt{k}}} + \sum_{k' = N} ^ {n} \frac{1}{2 ^ {\sqrt{k'}}} \leq \sum_{k = 1} ^ {N} \frac{1}{2 ^ {\sqrt{k}}} + \sum_{k' = N} ^ {n} \frac{1}{k' ^ 2}[/tex].

Can you go from here? :)
 
  • #5
[tex] \exists N=16 \forall n > N : n^4 < 2^n[/tex]
[tex] \forall n > N : \sum_{k = 1} ^ {n} \frac{1}{2 ^ {\sqrt{k}}} = \sum_{k = 1} ^ {N} \frac{1}{2 ^ {\sqrt{k}}} + \sum_{l = N+1} ^ {n} \frac{1}{2 ^ {\sqrt{l}}} < \sum_{k = 1} ^ {N} \frac{1}{2 ^ {\sqrt{k}}} + \sum_{l = N+1} ^ {n} \frac{1}{l ^ 2}<\sum_{k = 1} ^ {N} \frac{1}{2 ^ {\sqrt{k}}} + \sum_{l = N+1} ^ {n} \frac{1}{(l-1)l}=\sum_{k = 1} ^ {N} \frac{1}{2 ^ {\sqrt{k}}} + \sum_{l = N+1} ^ {n} \left(\frac{1}{l-1}-\frac{1}{l}\right)=[/tex]
[tex]=\sum_{k = 1} ^ {N} \frac{1}{2 ^ {\sqrt{k}}} + \frac{1}{N}-\frac{1}{n}<\sum_{k = 1} ^ {N} \frac{1}{2 ^ {\sqrt{k}}} + \frac{1}{N}[/tex]

So, increasing sequence has an upper bound, so it is convergent.
Q.E.D.

Is it correct ?
Thanks for help :-))) I am very happy now
 
Last edited:
  • #6
paniurelis said:
[tex] \exists N=16 \forall n > N : n^4 < 2^n[/tex]
[tex] \forall n > N : \sum_{k = 1} ^ {n} \frac{1}{2 ^ {\sqrt{k}}} = \sum_{k = 1} ^ {N} \frac{1}{2 ^ {\sqrt{k}}} + \sum_{l = N+1} ^ {n} \frac{1}{2 ^ {\sqrt{l}}} < \sum_{k = 1} ^ {N} \frac{1}{2 ^ {\sqrt{k}}} + \sum_{l = N+1} ^ {n} \frac{1}{l ^ 2}<\sum_{k = 1} ^ {N} \frac{1}{2 ^ {\sqrt{k}}} + \sum_{l = N+1} ^ {n} \frac{1}{l (l+1)}=\sum_{k = 1} ^ {N} \frac{1}{2 ^ {\sqrt{k}}} + \sum_{l = N+1} ^ {n} \left(\frac{1}{l}-\frac{1}{l+1}\right)=[/tex]
[tex]=\sum_{k = 1} ^ {N} \frac{1}{2 ^ {\sqrt{k}}} + \frac{1}{N+1}-\frac{1}{n+1}<\sum_{k = 1} ^ {N} \frac{1}{2 ^ {\sqrt{k}}} + \frac{1}{N+1}[/tex]

So, increasing sequence has an upper bound, so it is convergent.
Q.E.D.

Is it correct ?
Thanks for help :-))) I am very happy now

Yup, looks excellent. ^^! Well done. :biggrin:
 
  • #7
I have corrected error in my post above, was [tex]\frac{1}{k^2}<\frac{1}{k(k+1)}[/tex],
corrected [tex]\frac{1}{k^2}<\frac{1}{(k-1)k}[/tex]
 

What is proof of sequence convergence?

Proof of sequence convergence is a mathematical concept used to determine whether a sequence of numbers approaches a certain value as the number of terms in the sequence increases. It is an important tool in calculus and analysis, as it helps to understand the behavior of functions and their limits.

How is proof of sequence convergence different from proof of series convergence?

Proof of sequence convergence deals with individual terms in a sequence, while proof of series convergence deals with the sum of terms in a series. In other words, a sequence is a list of numbers, while a series is the sum of those numbers.

What are some common methods used to prove sequence convergence?

Some common methods used to prove sequence convergence include the squeeze theorem, the monotone convergence theorem, and the Cauchy criterion. Each method has its own set of assumptions and conditions, but they all ultimately determine whether a sequence approaches a certain value or not.

Can all sequences be proved to converge?

No, not all sequences can be proved to converge. Some sequences may diverge, meaning they do not approach a limit or a specific value. In these cases, other methods may be used to determine the behavior of the sequence, such as finding a pattern or using numerical methods.

How is proof of sequence convergence used in real-world applications?

Proof of sequence convergence has many real-world applications in fields such as physics, engineering, and economics. It is used to model and predict the behavior of systems and processes, such as the growth of a population or the decay of radioactive materials. It is also used in optimization problems to find the most efficient solution.

Similar threads

Replies
1
Views
1K
Replies
14
Views
1K
Replies
11
Views
970
Replies
11
Views
1K
Replies
18
Views
2K
Replies
1
Views
1K
Replies
5
Views
363
  • Calculus
Replies
13
Views
1K
Back
Top