Recursive sequences - notation question

In summary, the first sequence described recursively has a starting term of 2 and follows a pattern of each term being the square root of 3 times the previous term. The second sequence, which is described recursively as starting with an arbitrary positive number and each term being 6 divided by 2 times the previous term plus 1, has multiple possible interpretations and its behavior depends on the starting term. It can be shown that the sequence converges to a limit of 2, but the assumptions made about its monotonicity and boundedness in the process may not be accurate.
  • #1
marksyncm
100
5
I understand that when a sequence is described recursively, for example: ##a_1=2, a_{n+1} = \sqrt{3a_n}## then we mean that the first term is 2, the second term is ##\sqrt{3*2} = \sqrt{6}##, the third term is ##\sqrt{3*\sqrt{6}}##, and so on.

What I do not understand is how to interpret the following recursively described sequence: ##a_0>0, a_{n+1} = \frac{6}{2a_n+1}##. What is the first term of this sequence? Does ##a_0>0## mean that the first term can be any natural number? Or that it can be any real number?
 
Mathematics news on Phys.org
  • #2
Yes here ##a_0 > 0## means that ##a_0## is an arbitrary positive number (not necessarily natural though) that your sequence starts with. Your sequence will depend on this first term.
 
  • Like
Likes marksyncm
  • #3
Thank you.
 
  • #4
Decided to ask a follow-up question here rather than start a new thread, hope that's OK.

I'm reading up about finding limits of recursive functions. For example, ##a_1=\sqrt{2}, a_{n+1} = \sqrt{2a_n}##.

As far as I understand, the procedure is as follows:

1) ##\lim_{n \to \infty} a_{n+1} = \lim_{n \to \infty} \sqrt{2a_n} \rightarrow g = \sqrt{2g} \rightarrow g = -1## or ##4##. Because the sequence does not take on negative values, we can disregard ##-1## and focus on ##4##.

If the sequence converges to ##4##, then it must either be monotonically increasing and bounded from above or monotonically decreasing and bounded from below (please correct me if my understanding is incorrect/incomplete). This sequence appears to be increasing:

2) ##a_{n+1} > a_{n} \rightarrow \sqrt{2a_n} > a_n \rightarrow a_n(a_n-2) < 0##. This means that the sequence is monotonically increasing when the sequence ##a_n## is between ##0## and ##2##.

We now need to show that the sequence is bounded by ##2## from above:

3) ##a_{n+1}<2 \rightarrow \sqrt{2a_n} < 2 \rightarrow a_n < 2##

So the sequence is apparently bounded from above by ##2## and is monotonically increasing when it's smaller than ##2##, therefore it converges to ##2##.

However, what happens if, before point (2) above, I assume that the sequence is monotonically decreasing and bounded by ##2## from below? We then get that:

2b) ##a_{n+1} < a_{n} \rightarrow \sqrt{2a_n} < a_n \rightarrow a_n(a_n-2) > 0##. This means that the sequence is monotonically increasing when the sequence ##a_n## is larger than ##2## (or smaller than ##0##, but we disregard this).

We now show that the sequence is bounded by ##2## from below:

3b) ##a_{n+1} > 2 \rightarrow \sqrt{2a_n} > 2 \rightarrow a_n > 2##

I've (apparently) shown that the sequence is greater than ##2## and bounded by ##2## from below, which gives us the correct limit of the sequence even though it doesn't accurately describe what the sequence does / how it "behaves."

What am I doing wrong?

Thank you.
 
  • #5
marksyncm said:
1) limn→∞an+1=limn→∞√2an→g=√2g→g=−1\lim_{n \to \infty} a_{n+1} = \lim_{n \to \infty} \sqrt{2a_n} \rightarrow g = \sqrt{2g} \rightarrow g = -1 or 44. Because the sequence does not take on negative values, we can disregard −1-1 and focus on 44.

You might want to check your algebra. ##g = -1##?? Or ##g = 4##?

marksyncm said:
If the sequence converges to 44, then it must either be monotonically increasing and bounded from above or monotonically decreasing and bounded from below (please correct me if my understanding is incorrect/incomplete). This sequence appears to be increasing:

You have the implication the wrong way. If it is monotone increasing and bounded above or it is monotone descreasing and bound below, then it must converge.

Note that in general a convergent sequence need not do either of these things. It could oscillate either side of the limit.

marksyncm said:
2) an+1>an→√2an>an→an(an−2)<0a_{n+1} > a_{n} \rightarrow \sqrt{2a_n} > a_n \rightarrow a_n(a_n-2) < 0. This means that the sequence is monotonically increasing when the sequence ana_n is between 00 and 22.

Again, the implication is the wrong way round. You really want to show that

##0 < a_n < 2 \ \Rightarrow \ 0 < a_n < a_{n+1} < 2##

marksyncm said:
We now need to show that the sequence is bounded by 22 from above:

3) an+1<2→√2an<2→an<2

Again, this is going backwards.

marksyncm said:
However, what happens if, before point (2) above, I assume that the sequence is monotonically decreasing and bounded by 22 from below? We then get that:

2b) an+1<an→√2an<an→an(an−2)>0a_{n+1} < a_{n} \rightarrow \sqrt{2a_n} < a_n \rightarrow a_n(a_n-2) > 0. This means that the sequence is monotonically increasing when the sequence ana_n is larger than 22 (or smaller than 00, but we disregard this).

We now show that the sequence is bounded by 22 from below:

3b) an+1>2→√2an>2→an>2a_{n+1} > 2 \rightarrow \sqrt{2a_n} > 2 \rightarrow a_n > 2

I've (apparently) shown that the sequence is greater than 22 and bounded by 22 from below, which gives us the correct limit of the sequence even though it doesn't accurately describe what the sequence does / how it "behaves."

What am I doing wrong?

You tried to show that if ##a_n > 2## then the sequence is monotone decreasing and bounded below by 2. But, again, the implications are the wrong way round.

In summary, the behaviour of this sequence depends on where it starts, i.e. on ##a_1##., and you can get one of the two options.
 
  • #6
Thank you for the response.

PeroK said:
Again, the implication is the wrong way round. You really want to show that

##0 < a_n < 2 \ \Rightarrow \ 0 < a_n < a_{n+1} < 2##

I am not sure I follow. Could you show what you mean here? I can't seem to find a way to show that ##0 < a_n < 2## without first showing that ##a_{n+1} < 2## (because the ##a_n## term is "hidden" under the square root term that describes ##a_{n+1}##.
 
  • #7
marksyncm said:
Thank you for the response.
I am not sure I follow. Could you show what you mean here? I can't seem to find a way to show that ##0 < a_n < 2## without first showing that ##a_{n+1} < 2## (because the ##a_n## term is "hidden" under the square root term that describes ##a_{n+1}##.

That piece of logic does not show that ##0 < a_n < 2##. It says that IF ##0 < a_n < 2## THEN ##0 < a_{n} < a_{n+1} < 2##.

The whole thing then hinges on the value of ##a_1##. If ##0 < a_1 < 2##, then by induction you have:

##0 < a_1 < a_2 < a_3 \dots < 2##
 
  • #8
PeroK said:
It says that IF ##0 < a_n < 2##

Sorry for being dense, but doesn't this mean that we need to find out IF ##0 < a_n < 2##? Because ##0 < a_n < a_{n+1} < 2## seems to follow from ##0 < a_n < 2## (isn't that what implication means? IF ##A## THEN ##B##, meaning we want to check IF ##A## because if it is, then ##B##?), then don't we first need to show that ##0 < a_n < 2##?
 
  • #9
marksyncm said:
Sorry for being dense, but doesn't this mean that we need to find out IF ##0 < a_n < 2##? Because ##0 < a_n < a_{n+1} < 2## seems to follow from ##0 < a_n < 2## (isn't that what implication means? IF ##A## THEN ##B##, meaning we want to check IF ##A## because if it is, then ##B##?), then don't we first need to show that ##0 < a_n < 2##?

This is primarily a matter of logic. The conclusion follows logically from the premise. You don't have to verify the premise for the logic to hold. Let's say I have proved that:

If ##0 < a_n < 2##, then ##0 < a_n < a_{n+1} < 2##.

Then, by induction, I can prove that:

If ##0 < a_1 < 2##, then the sequence ##\{a_n\}## converges to ##2##.

At that point the mathematical work is done. I can sit down and have a cup of tea.

If someone comes along and tells me they have this recursive sequence and ##a_1 = 1.5##, then I know that the sequence converges to ##2##. If instead they tell me that ##a_1 = 3##, then I have to put down my tea and do some more work. Because I haven't analysed that case yet.

You need to separate in your mind the logical sequence of steps, which lead to proofs, lemmas, theorems from the application of that logic when you have the specific case.
 
  • Like
Likes marksyncm
  • #10
PeroK said:
If ##0 < a_n < 2##, then ##0 < a_n < a_{n+1} < 2##.

To make sure, in this part above, ##a_n## refers to any (and all) term of the sequence, and ##a_{n+1}## refers to the term that follows it?
 
  • #11
marksyncm said:
To make sure, in this part above, ##a_n## refers to any (and all) term of the sequence, and ##a_{n+1}## refers to the term that follows it?

Yes. The only thing you need is that ##a_{n+1} = \sqrt{2a_n}##. Because this relationship holds for all ##n##, then the logic holds for all ##n##. The rest is induction, depending on ##a_1##.
 

FAQ: Recursive sequences - notation question

What is a recursive sequence?

A recursive sequence is a sequence of numbers where the next term is determined by a function or rule applied to the previous term. In other words, the formula for each term in the sequence contains the previous term. For example, the Fibonacci sequence (1, 1, 2, 3, 5, 8, 13, ...) is a well-known recursive sequence where each term is the sum of the two previous terms.

What is the notation used for recursive sequences?

The notation used for recursive sequences is typically an explicit formula, which uses a variable n to represent the term number, and a recursive formula, which uses a variable n-1 to represent the previous term. For example, the explicit formula for the Fibonacci sequence is Fn = Fn-1 + Fn-2, while the recursive formula is F1 = F2 = 1, and Fn = Fn-1 + Fn-2 for n > 2.

How do you write a recursive sequence in summation notation?

To write a recursive sequence in summation notation, you can use the sigma notation, where the starting term is represented by the lower limit of the sigma (usually 1), the ending term is represented by the upper limit of the sigma, and the formula for each term is written after the sigma symbol. For example, the Fibonacci sequence can be written in summation notation as Σ n=1 to ∞ Fn = 1 + 1 + 2 + 3 + 5 + 8 + 13 + ...

Can recursive sequences have negative terms?

Yes, recursive sequences can have negative terms. The formula for each term can include negative numbers, and the sequence can continue to alternate between positive and negative terms. For example, the sequence (-1, 3, -5, 7, -9, ...) is a recursive sequence where each term is the previous term multiplied by -2 and the first term is -1.

How do you determine the pattern of a recursive sequence?

The pattern of a recursive sequence can be determined by examining the formula for each term and identifying how the previous term is used to calculate the next term. You can also look at the differences between consecutive terms to see if there is a consistent pattern. For example, the sequence (1, 4, 9, 16, 25, ...) has a pattern of adding consecutive odd numbers (3, 5, 7, 9, ...), which is reflected in the formula Fn = Fn-1 + 2n - 1.

Similar threads

Replies
1
Views
1K
Replies
4
Views
3K
Replies
4
Views
1K
Replies
15
Views
1K
Replies
1
Views
1K
Replies
11
Views
2K
Replies
4
Views
1K
Replies
8
Views
2K
Back
Top