3

I've tried to prove it aiming for a contradiction:If $\frac{68}{25}$ is not an upper bound of A, then, $\exists n\in\Bbb N^*$ such that $$\frac{68}{25}\lt(1+\frac{1}{n})^n\Rightarrow\frac{(68)^{1/n}}{(25)^{1/n}}-\frac{1}{n}<1\Rightarrow n(68)^{1/n}-(25)^{1/n}\lt n(25)^{1/n}\Rightarrow n(68)^{1/n}\lt(25)^{1/n}(n+1)$$ At this part i get stuck, because, we see that $n<n+1$ and instinctively i'd say that

$(68)^{1/n}<(25)^{1/n}$, apparently, i´ve found a contradiction: $68<25$. But i've found that in general if

we have $a<b$ and $ac<bd$ it´s not always true that $c<d$, for instance, $2<3$ and $2\sqrt2<3(1)$

don't imply that $\sqrt2<1$. What can i do?, Please, help me. THANKS

orlp
  • 10,508
Tom Arbuckle
  • 113
  • 10

4 Answers4

5

The expression $\left(1 + \frac{1}{n}\right)^n$ is increasing, and its limit as $n \to \infty$ is $$e = \sum_{k = 0}^{\infty} \frac{1}{k!} ,$$ so as hinted in the comments, the problem is equivalent to showing that $$ e \leq \frac{68}{25}.$$

Method 1: Taylor expansion The above series converges rapidly, suggesting we can produce tight bounds by keeping the first several terms and then bounding the tail by a small quantity.

Expanding the series, we have $$e = 1 + 1 + \frac{1}{2} + \frac{1}{6} + \frac{1}{24} + \frac{1}{120} + \frac{1}{720} + R ,$$ for a remainder series $R$. But the $\ell$th term of $R$ is a product of $\frac{1}{720}$ with a product of $\ell$ integers $> 2$, so $$R < \frac{1}{720} \sum_{\ell = 1}^{\infty} 2^{-\ell} = \frac{1}{720}.$$ Combining this with the previous inequality gives $e < \frac{979}{630}$, and cross-multipliying gives $\frac{979}{630} < \frac{68}{25}$. Of course, a similar method could be applied to other series with positive terms converging to $e$.

Method 2: Continued fraction expansion From the series characterization of $e$, its continued fraction expansion is $[2; 1, 2, 1, 1, 4, 1, 1, 6, \ldots, 1, 1, 2n, \ldots]$. The successive convergents of a continued fraction expansions (1) approximate the value more and more closely and (2) alternate being underestimates and overestimates. Computing from the expansion gives that the convergents of this series are: $2, 3, \frac{8}{3}, \frac{11}{4}, \frac{19}{7}, \frac{87}{32}, \ldots .$ Since $\frac{19}{7} < \frac{87}{32}$, the latter is an overestimate for $e$, and again cross-multiplying gives $e < \frac{87}{32} < \frac{68}{25}$.

Travis Willse
  • 99,363
2

Approach 1

As shown in this answer, $\left(1+\frac1n\right)^n$ is increasing and $\left(1+\frac1n\right)^{n+1}$ is decreasing.

Note that $\left(1+\frac1{791}\right)^{792}=2.71999990408\lt2.72=\frac{68}{25}$

For $n\le791$, $\left(1+\frac1n\right)^n\le\left(1+\frac1{791}\right)^{791}\lt\left(1+\frac1{791}\right)^{792}\lt\frac{68}{25}$

For $n\ge791$, $\left(1+\frac1n\right)^n\lt\left(1+\frac1n\right)^{n+1}\le\left(1+\frac1{791}\right)^{792}\lt\frac{68}{25}$


Approach 2

Since $\left(1+\frac1n\right)^n$ is increasing and, as shown in this answer, $$ \lim\limits_{n\to\infty}\left(1+\frac1n\right)^n=\sum_{n=0}^\infty\frac1{n!} $$ it follows from the inequality, $$ \begin{align} \sum_{n=0}^\infty\frac1{n!} &=\frac1{0!}+\frac1{1!}+\frac1{2!}+\frac1{3!}+\frac1{4!}\sum_{k=4}^\infty\frac{4!}{k!}\\ &\le\frac1{0!}+\frac1{1!}+\frac1{2!}+\frac1{3!}+\frac1{4!}\sum_{k=4}^\infty\frac1{5^{k-4}}\\ &=\frac1{0!}+\frac1{1!}+\frac1{2!}+\frac1{3!}+\frac1{4!}\frac54\\[6pt] &=\frac{87}{32}\\[6pt] &\lt\frac{68}{25} \end{align} $$ that $\left(1+\frac1n\right)^n\lt\frac{68}{25}$.

robjohn
  • 345,667
  • 2
    Perhaps the idea of your answer is the same, but you left out any justification that the left sequence was increasing and that the right sequence is decreasing, and once you do that, there are two cases to consider. So I don't believe this answer is a duplicate of yours. – robjohn Nov 01 '18 at 14:25
  • And the second approach is similar to the first approach in mine, but robjohn's method is better in the sense that one doesn't need to deal with three-digit numbers and so saves a little bit of time when cross-multiplying to verify the inequality. – Travis Willse Nov 01 '18 at 18:09
1

We have $$\left(1+\dfrac1n\right)^n = \sum_{k=0}^n \dbinom{n}k \dfrac1{n^k} \leq \sum_{k=0}^n \dfrac1{k!} = 1 + \dfrac1{1!} + \dfrac1{2!} + \dfrac1{3!} + \dfrac1{4!} + \sum_{k=5}^n \dfrac1{k!}$$ For $k \geq 5$, we have $$k! \geq 5! \times 5^{k-5} = 120 \times 5^{k-5}$$ Hence, $$\sum_{k=5}^n \dfrac1{k!} \leq \dfrac1{120} \sum_{k=5}^{\infty} \dfrac1{5^{k-5}} = \dfrac1{120} \times \dfrac54 = \dfrac1{96}$$ Hence, we see that $$\left(1+\dfrac1n\right)^n \leq 1 + \dfrac1{1!} + \dfrac1{2!} + \dfrac1{3!} + \dfrac1{4!} + \dfrac1{96} = 2+\dfrac{48+16+4+1}{96} = 2 + \dfrac{23}{32} \leq 2+\dfrac{18}{25}$$

Adhvaitha
  • 20,259
-2

$$ \left(1+\frac1n\right)^n < \left(1+\frac1n\right)^{n+1} $$ and the right hand side is decreasing. From this you get $\sup A < \left(1+\frac1{1000}\right)^{1000+1}<\frac{68}{25}$.

Edit for the non-believers: the last inequality is a true fact which can be verified with exact arithmetic: $25\times1001^{1001}<68\times1000^{1001}$.

I still don't get how this correct answer can get downvoted (but still get also copied by later answers...).

Federico
  • 5,900
  • Second inequality not proven. Right hand side decreasing????? – herb steinberg Nov 01 '18 at 01:53
  • @herbsteinberg You can compute and check the second inequality. And yes, the right hand side is a decreasing function. – Federico Nov 01 '18 at 02:25
  • And why the downvote?! This answer is both correct and simple. – Federico Nov 01 '18 at 02:28
  • $(1+\frac{1}{n})^n$ is an increasing function of $n$, so how does multiplying by $(1+\frac{1}{n})$ make it decreasing? – herb steinberg Nov 01 '18 at 02:30
  • @herbsteinberg $1+1/n$ is a decreasing function. you are multiplying an increasing and a decreasing function. The result can be decreasing. So what? – Federico Nov 01 '18 at 02:34
  • If you don't trust me and are unable to prove it: https://www.wolframalpha.com/input/?i=plot+(1+%2B+1%2Fn)%5E(n+%2B+1)+for+n+from+1+to+10 – Federico Nov 01 '18 at 02:34
  • And please, explain the downvote – Federico Nov 01 '18 at 02:35
  • $(1+\frac{1}{n})^n$ is an increasing function of $n$. You claim it is decreasing and you are wrong! For $n=1,2,3,4$, the expression $=2,2.25,2.37..,2.44..$. You could get a formal proof by using a binomial expansion and see that term by term it is increasing and the limit is the power series for $e$. – herb steinberg Nov 01 '18 at 02:45
  • $(1+1/n)^n$ is increasing, $(1+1/n)^{n+1}$ is decreasing – Federico Nov 01 '18 at 02:48
  • When did ever say that $(1+1/n)^n$ is decreasing?!? I said the right hand side – Federico Nov 01 '18 at 02:51
  • You need to prove it! – herb steinberg Nov 01 '18 at 02:52
  • You need to remove the downvote if you want me to do your homework. – Federico Nov 01 '18 at 02:53
  • Can you at least search on this site? https://math.stackexchange.com/questions/2071492/the-function-f-n-1-1-n-n1-is-decreasing – Federico Nov 01 '18 at 02:54
  • I'm not doing any homework! The site seems to show your assertion is correct. Your last line needs further verification. Did you give my answer a down vote - if so why? – herb steinberg Nov 01 '18 at 03:03
  • @Federico If you're in need of light relief after this assault, have a look at the oddly synchronistic nonsense in some of the comments on this very recent question, as well. It's embarrassing to yours truly! There must be something in the air - that's my excuse, and I'm sticking to it. :) – Calum Gilhooley Nov 01 '18 at 13:58
  • @CalumGilhooley How was my comment a "slur"? I just pointed out that part 1 of his (later) answer was very heavily inspired by mine, to say the least... In the meantime, I had to also waste time arguing with herb over some nonsense. If anything, I might be more in a mood to remove my contribution to this question altogether, since for some inexplicable reason I also managed to score a gratifying -2. By the way, Jack's answer with magical integration is nice as always. – Federico Nov 02 '18 at 18:54
  • I don't know why your answer was badly received. It would be a shame if you feel pressured into deleting it, by the downvotes, and by all that arguing. – Calum Gilhooley Nov 02 '18 at 19:06
  • @CalumGilhooley I'm in peace with humanity again :) I removed the alleged "slur" and I will leave this shameful answer alone. Have a nice day computing $1001^{1001}$ by hand :D – Federico Nov 02 '18 at 19:14
  • @Federico Don't think I won't do it! :) – Calum Gilhooley Nov 02 '18 at 19:40
  • @Federico Excuse me if this is stupid, but, why do we need to know that $(1+\frac{1}{n})^n$ is increasing and $(1+\frac{1}{n})^{n+1}$ is decreasing?. I mean why is not enough to know that for any natural n, $(1+\frac{1}{n})^n$<$(1+\frac{1}{n})^{n+1}$ ? I hope you can answer this and don´t get angry. – Tom Arbuckle Dec 13 '18 at 01:20
  • @TomArbuckle becouse I'm using both $\max_{n\leq1000}(1+1/n)^n\leq(1+1/1000)^{1000} < (1+1/1000)^{1001} < 68/25$ and $\sup_{n\geq1000}(1+1/n)^n\leq\sup_{n\geq1000}(1+1/n)^{n+1}\leq(1+1/1000)^{1001} < 68/25$ – Federico Dec 13 '18 at 14:24
  • In the first I use $(1+1/n)^n$ increasing, in the second $(1+1/n)^{n+1}$ decreasing – Federico Dec 13 '18 at 14:26
  • If you only know $(1+1/n)^n<(1+1/n)^{n+1}$, all you can say is $\sup_n (1+1/n)^n\leq\sup_n(1+1/n)^{n+1}=(1+1/1)^{2}=4$ – Federico Dec 13 '18 at 14:27