7 − 7 − 7 − 7 − 7 − … 1 2 1 2 1 2 1 2 = ?
This section requires Javascript.
You are seeing this because something didn't load right. We suggest you, (a) try
refreshing the page, (b) enabling javascript if it is disabled on your browser and,
finally, (c)
loading the
non-javascript version of this page
. We're sorry about the hassle.
Great! Thanks for explaining why this sequence converges to the value of 4. Note that it is not always guaranteed that the limit must exist and be finite.
I came up with a similar solution and posted it on Facebook. A couple people have challenged it, saying that there's no reason that you can't choose a different seed value a 1 . This challenge makes sense, as the seed value gets buried deep down infinitely far in the fraction anyway. So for instance you could choose a 1 = 5 , and still you find that the sequence converges to 4.
The problem starts when you consider a 1 = 3 . When you use this seed value the sequence converges to 3, and there's nothing stopping you from using this seed value.
How would you respond to this? My contention is that the use of the ellipsis in the problem statement means "continue in this pattern", which to me means that a 1 must be 7.
Log in to reply
The dynamic system x n + 1 = 7 − 1 2 / x n has two fixed points, namely 3 and 4; however, 3 is unstable and repels, while 4 is stable and attracts.
Proof: let f ( x ) = 7 − 1 2 / x . Then the derivative is f ′ ( x ) = 1 2 / x 2 .
Since f ′ ( 3 ) = 1 2 / 3 2 = 3 4 > 1 , we have the following behavior near 3: if x n − 3 = ε , then x n + 1 − 3 = 3 4 ε > ε , showing that any sequence starting near 3 will move away from it.
On the other hand, f ′ ( 4 ) = 1 2 / 4 2 = 4 3 < 1 . If x n − 4 = ε , then x n + 1 − 4 = 4 3 ε < ε , showing that any sequence starting near 4 will get closer to 4.
Further analysis shows the following behavior:
if we start with x > 4 , the sequence decreases monotonously toward 4.
if we start with 3 < x < 4 , the sequence increases monotonously toward 4.
if we start with 7 1 2 < x < 3 , the sequence decreases monotonously, until ...
if we start with 0 < x < 7 1 2 , the next step gets us to ...
if we start with x < 0 , the next step gives x > 7 , and from there the sequence decreases monotonously toward 4.
Log in to reply
Thank you. I just posted a completely different solution which I think avoids the problem of arbitrarily chosen seed values.
Indeed, that's true that the seed value is "hidden" in the ellipsis, if we assume the recurrence relation to mean building the fraction upward.
However, there's a different interpretation. Like pretty much all continued fractions, the sequence generated is actually the sequence of convergents, cutting off after several levels. Thus, for example,
1 + 2 + 3 + 4 + … 1 1 1
is the limit of the sequence:
⎝ ⎜ ⎜ ⎜ ⎜ ⎜ ⎛ 1 , 1 + 2 1 , 1 + 2 + 3 1 1 , 1 + 2 + 3 + 4 1 1 1 , … ⎠ ⎟ ⎟ ⎟ ⎟ ⎟ ⎞
For this particular case, the sequence is:
⎝ ⎜ ⎜ ⎜ ⎜ ⎜ ⎛ 7 , 7 − 7 1 2 , 7 − 7 − 7 1 2 1 2 , 7 − 7 − 7 − 7 1 2 1 2 1 2 , … ⎠ ⎟ ⎟ ⎟ ⎟ ⎟ ⎞
which just so happens to also be generated by the given recurrence relation above as well. This sequence of convergents definition is also easily generalizable, while building bottom-up won't give anything for the [ 1 ; 2 , 3 , 4 , … ] example above. Thus I believe choosing a 1 = 7 is justified.
Log in to reply
I agree that a 1 = 7 is justified. Just the same, I have posted a solution here that I think avoids the problem altogether. I defined a different sequence in which the seed values show up at the top of the continued fraction and cannot be arbitrary. I also posted it on Facebook, and I hope it satisfies the critics. They are saying that this continued fraction has no unique value, which I think is dead wrong.
I'm using the continued fractions and I found 2 possible answer, just like the name of this post lol
Log in to reply
That's your job to figure out which, if any, of them is the actual answer.
I didn't understand how you ruled out 3, please elaborate
Log in to reply
The epsilon-delta definition of limit states that a sequence { a n } converges to a limit L exactly if for every ϵ > 0 , there exists N such that for all n > N , ∣ a n − L ∣ < ϵ .
We claim that L = 3 is impossible. Suppose that L = 3 is possible, then we can plug it into the definition above: for every ϵ > 0 , there exists N such that for all n > N , ∣ a n − 3 ∣ < ϵ . But now consider ϵ = 1 . We proved that a n > 4 , which means a n − 3 > 1 . In particular, since a n − 3 > 1 > 0 , this is positive, so ∣ a n − 3 ∣ = a n − 3 > 1 for all n . Thus it's impossible to obtain such N . This is a contradiction, so the assumption that L = 3 is false; that is, the limit is not 3.
Log in to reply
This argument is not convincing. If we start with a 1 = 3 (and why not?), we get a n = 3 for all subsequent values of n . Thus for every ε > 0 , and for all n > N = 1 , it is true that ∣ a n − 3 ∣ = 0 < ε , showing that the limit in this case is 3.
Log in to reply
@Arjen Vreugdenhil – Choosing the seed value is also a problem, but I believe a 1 = 7 is "intuitive". I also gave a different argument in a comment elsewhere on this page that a 1 = 7 is "right".
7x-12=x^2 then two values formed 3 and 4..
Log in to reply
@Sudhanshu Mishra – That's already mentioned in the solution. The problem is now to determine whether the answer is 3, 4, or none of them (the fraction doesn't converge). My solution above shows that it's 4.
I literally don't know how I solved it I just came up with the answer :)
To save myself some typing, I'll tell you that I'm using the formalism of the following article. Generalized Continued Fraction
In this problem, a n = − 1 2 and b n = 7 for all n . Then A n = 7 A n − 1 − 1 2 A n − 2 and B n = 7 B n − 1 − 1 2 B n − 2 . The seed values for these relations are as follows.
A 0 = b 0 = 7 , A 1 = b 0 b 1 + a 1 = 3 7 , B 0 = 1 , B 1 = b 1 = 7
We wish to solve the recurrence relations for A n and B n . The standard trick for solving linear recurrence relations such as these is to solve the characteristic polynomial equation for the relation.
The characteristic equation for A n is x 2 = 7 x − 1 2 , and its solutions are x = 3 or x = 4 . So the solution of the recurrence relation is of the form A n = c 1 3 n + c 2 4 n . Using the seed values, we obtain c 1 = − 9 , c 2 = 1 6 , so the solution is A n = − 9 ⋅ 3 n + 1 6 ⋅ 4 n .
B n has the same characteristic equation, so B n = d 1 3 n + d 2 4 n . Using the seed values, we obtain d 1 = − 3 , d 2 = 4 , so the solution is B n = − 3 ⋅ 3 n + 4 ⋅ 4 n .
Now define a sequence x n as follows.
x n = B n A n = − 3 ⋅ 3 n + 4 ⋅ 4 n − 9 ⋅ 3 n + 1 6 ⋅ 4 n
Dividing the numerator and denominator by 4 n yields the following.
x n = − 3 ( 4 3 ) n + 4 − 9 ( 4 3 ) n + 1 6
Finally, take the limit of the sequence.
lim n → ∞ x n = lim n → ∞ − 3 ( 4 3 ) n + 4 − 9 ( 4 3 ) n + 1 6 = − 3 ⋅ 0 + 4 − 9 ⋅ 0 + 1 6 = 4
Problem Loading...
Note Loading...
Set Loading...
First, we need to formalize what it means by
7 − 7 − 7 − … 1 2 1 2 1 2
The common definition is this: define the sequence a 1 = 7 , a n + 1 = 7 − a n 1 2 , then the expression above is defined to be n → ∞ lim a n .
If the sequence converges to a limit L , then it satisfies L = 7 − L 1 2 . But there are two solutions, L = 3 and L = 4 . Which one is correct? Moreover, it's still possible that neither of them is correct, if the sequence doesn't converge.
This is a pretty common problem; I posted a solution to a similar problem . The steps are generally in this form:
For this problem, we will follow the same steps:
Step 1 : Find the candidates
This was done above.
Step 2 : Sequence is bounded below by 4
We proceed by induction. Note that a 1 = 7 > 4 , so this serves as the base case. For the inductive step, note that since a n > 4 , we have a n 1 2 < 3 , and so a n + 1 = 7 − a n 1 2 > 7 − 3 = 4 .
Step 3 : This rules out 3 as a limit
We use the epsilon-delta definition of limit for this. A sequence { a n } converges to a limit L precisely when for every ϵ > 0 , there exists N > 0 such that for all n > N , we have ∣ a n − L ∣ < ϵ .
Suppose, for the sake of contradiction, that L = 3 . Since the sequence is bounded below by 4, we have a n > 4 for all n . This means a n − 3 > 1 . Since a n − 3 > 1 > 0 , a n − 3 is positive, so ∣ a n − 3 ∣ = a n − 3 . Now consider ϵ = 1 . We have ∣ a n − L ∣ = ∣ a n − 3 ∣ = a n − 3 > 1 for all n , contradiction with the requirement that there exists some N such that ∣ a n − L ∣ < 1 for all n > N . Thus the assumption that L = 3 is incorrect; 3 cannot be the limit.
Step 4 : The sequence is monotonically decreasing
This needs the use of Step 2. Since a n > 4 , we know that ( a n − 4 ) ( a n − 3 ) > 0 ⋅ 1 > 0 , so it is positive. We can rearrange that into a n 2 − 7 a n + 1 2 > 0 or a n 2 > 7 a n − 1 2 . Since a n > 4 , a n is positive, so dividing by a n doesn't flip the inequality sign: a n > 7 − a n 1 2 = a n + 1 . This completes the proof.
Step 5 : The sequence converges
We apply the monotone convergence theorem. The sequence is monotonically decreasing, and it is bounded below, so it converges.
Step 6 : The sequence converges to 4
We know, from Step 1, that if the sequence converges, then it converges to either 3 or 4. From Step 5, we know that the sequence does converge. From Step 3, the limit cannot be 3. The only remaining option is that the limit is 4; that is, the value of the expression above is 4 .