Limit of continued fraction Announcing the arrival of Valued Associate #679: Cesar Manara Planned maintenance scheduled April 17/18, 2019 at 00:00UTC (8:00pm US/Eastern)A new continued fraction for Apéry's constant, $zeta(3)$?Continued fraction and double series.Why are there no continued fraction representation for $pi$ obeying mathematical rules?a certain simple continued fractionconjectured general continued fraction for the quotient of gamma functionscontinued fraction $F(x)$ that is a generating function of central binomial coefficientsIs an arbitrary continued fraction convergent?How to find other Ramanujan-type continued fractionsUsing continued fraction convergents to find a formula in terms of derivativesConjectured continued fraction for the Generalized Rogers-Ramanujan continued fraction
When is phishing education going too far?
Why is "Captain Marvel" translated as male in Portugal?
Stars Make Stars
How do I stop a creek from eroding my steep embankment?
Is there a concise way to say "all of the X, one of each"?
How can players work together to take actions that are otherwise impossible?
"Seemed to had" is it correct?
Are my PIs rude or am I just being too sensitive?
Why don't the Weasley twins use magic outside of school if the Trace can only find the location of spells cast?
Is above average number of years spent on PhD considered a red flag in future academia or industry positions?
3 doors, three guards, one stone
Is the address of a local variable a constexpr?
Do I really need recursive chmod to restrict access to a folder?
Why does Python start at index -1 when indexing a list from the end?
Why aren't air breathing engines used as small first stages
Why was the term "discrete" used in discrete logarithm?
Disable hyphenation for an entire paragraph
List *all* the tuples!
What is the musical term for a note that continously plays through a melody?
Output the ŋarâþ crîþ alphabet song without using (m)any letters
What is the longest distance a 13th-level monk can jump while attacking on the same turn?
The logistics of corpse disposal
Is it ethical to give a final exam after the professor has quit before teaching the remaining chapters of the course?
Center align columns in table ignoring minus signs?
Limit of continued fraction
Announcing the arrival of Valued Associate #679: Cesar Manara
Planned maintenance scheduled April 17/18, 2019 at 00:00UTC (8:00pm US/Eastern)A new continued fraction for Apéry's constant, $zeta(3)$?Continued fraction and double series.Why are there no continued fraction representation for $pi$ obeying mathematical rules?a certain simple continued fractionconjectured general continued fraction for the quotient of gamma functionscontinued fraction $F(x)$ that is a generating function of central binomial coefficientsIs an arbitrary continued fraction convergent?How to find other Ramanujan-type continued fractionsUsing continued fraction convergents to find a formula in terms of derivativesConjectured continued fraction for the Generalized Rogers-Ramanujan continued fraction
$begingroup$
Let $(c_n)$ be a sequence of positive numbers converging to $1$ as $ntoinfty$. That is, $c_n > 0$ for every $n$ and $limlimits_ntoinftyc_n=1$. Let $g_1=-c_1$ and define, for $n geq 1$,$$g_n+1=frac-1c_n-g_n.$$
Question: Does $limlimits_n to infty g_n$ exist?
Notice that in case $g_n$ converges to, say $g_infty$, the limit is given by the negative root of the quadratic equation determined by the equation $g_infty = dfrac-11-g_infty$.
Notice also that each $g_n+1$ can be represented as the following finite continued fraction
$$g_n+1=-cfrac1c_n+cfrac1c_n-1+cfrac1ddots+cfrac1c_1-g_1,$$
the shorthand for which is sometimes given by the notation $g_n+1=[c_n,c_n-1,cdots, c_1-g_1]$.
The question is: Does $limlimits_n to infty g_n$ exist?
real-analysis sequences-and-series continued-fractions
$endgroup$
add a comment |
$begingroup$
Let $(c_n)$ be a sequence of positive numbers converging to $1$ as $ntoinfty$. That is, $c_n > 0$ for every $n$ and $limlimits_ntoinftyc_n=1$. Let $g_1=-c_1$ and define, for $n geq 1$,$$g_n+1=frac-1c_n-g_n.$$
Question: Does $limlimits_n to infty g_n$ exist?
Notice that in case $g_n$ converges to, say $g_infty$, the limit is given by the negative root of the quadratic equation determined by the equation $g_infty = dfrac-11-g_infty$.
Notice also that each $g_n+1$ can be represented as the following finite continued fraction
$$g_n+1=-cfrac1c_n+cfrac1c_n-1+cfrac1ddots+cfrac1c_1-g_1,$$
the shorthand for which is sometimes given by the notation $g_n+1=[c_n,c_n-1,cdots, c_1-g_1]$.
The question is: Does $limlimits_n to infty g_n$ exist?
real-analysis sequences-and-series continued-fractions
$endgroup$
$begingroup$
Express $c_n$ from the recurrance formula: $c_n=g_n- frac1g_n+1$ taking the limits of both sides we have: $limlimits_nrightarrow infty(g_n- frac1g_n+1)=1tag1$ if $limlimits_nrightarrow inftyg_n=infty$ then (1) would not be valid. So the $limlimits_nrightarrow inftyg_n$ exists.
$endgroup$
– JV.Stalker
Mar 26 at 11:46
$begingroup$
Thanks for the comment. But doesn't this only prove that $lim_n to infty g_n$ does not equal infinity? Why does this prove that the limit $lim g_n$ exists?
$endgroup$
– Shibi Vasudevan
Mar 26 at 12:01
add a comment |
$begingroup$
Let $(c_n)$ be a sequence of positive numbers converging to $1$ as $ntoinfty$. That is, $c_n > 0$ for every $n$ and $limlimits_ntoinftyc_n=1$. Let $g_1=-c_1$ and define, for $n geq 1$,$$g_n+1=frac-1c_n-g_n.$$
Question: Does $limlimits_n to infty g_n$ exist?
Notice that in case $g_n$ converges to, say $g_infty$, the limit is given by the negative root of the quadratic equation determined by the equation $g_infty = dfrac-11-g_infty$.
Notice also that each $g_n+1$ can be represented as the following finite continued fraction
$$g_n+1=-cfrac1c_n+cfrac1c_n-1+cfrac1ddots+cfrac1c_1-g_1,$$
the shorthand for which is sometimes given by the notation $g_n+1=[c_n,c_n-1,cdots, c_1-g_1]$.
The question is: Does $limlimits_n to infty g_n$ exist?
real-analysis sequences-and-series continued-fractions
$endgroup$
Let $(c_n)$ be a sequence of positive numbers converging to $1$ as $ntoinfty$. That is, $c_n > 0$ for every $n$ and $limlimits_ntoinftyc_n=1$. Let $g_1=-c_1$ and define, for $n geq 1$,$$g_n+1=frac-1c_n-g_n.$$
Question: Does $limlimits_n to infty g_n$ exist?
Notice that in case $g_n$ converges to, say $g_infty$, the limit is given by the negative root of the quadratic equation determined by the equation $g_infty = dfrac-11-g_infty$.
Notice also that each $g_n+1$ can be represented as the following finite continued fraction
$$g_n+1=-cfrac1c_n+cfrac1c_n-1+cfrac1ddots+cfrac1c_1-g_1,$$
the shorthand for which is sometimes given by the notation $g_n+1=[c_n,c_n-1,cdots, c_1-g_1]$.
The question is: Does $limlimits_n to infty g_n$ exist?
real-analysis sequences-and-series continued-fractions
real-analysis sequences-and-series continued-fractions
edited Mar 29 at 16:36
Saad
20.6k92452
20.6k92452
asked Mar 26 at 9:28
Shibi VasudevanShibi Vasudevan
705617
705617
$begingroup$
Express $c_n$ from the recurrance formula: $c_n=g_n- frac1g_n+1$ taking the limits of both sides we have: $limlimits_nrightarrow infty(g_n- frac1g_n+1)=1tag1$ if $limlimits_nrightarrow inftyg_n=infty$ then (1) would not be valid. So the $limlimits_nrightarrow inftyg_n$ exists.
$endgroup$
– JV.Stalker
Mar 26 at 11:46
$begingroup$
Thanks for the comment. But doesn't this only prove that $lim_n to infty g_n$ does not equal infinity? Why does this prove that the limit $lim g_n$ exists?
$endgroup$
– Shibi Vasudevan
Mar 26 at 12:01
add a comment |
$begingroup$
Express $c_n$ from the recurrance formula: $c_n=g_n- frac1g_n+1$ taking the limits of both sides we have: $limlimits_nrightarrow infty(g_n- frac1g_n+1)=1tag1$ if $limlimits_nrightarrow inftyg_n=infty$ then (1) would not be valid. So the $limlimits_nrightarrow inftyg_n$ exists.
$endgroup$
– JV.Stalker
Mar 26 at 11:46
$begingroup$
Thanks for the comment. But doesn't this only prove that $lim_n to infty g_n$ does not equal infinity? Why does this prove that the limit $lim g_n$ exists?
$endgroup$
– Shibi Vasudevan
Mar 26 at 12:01
$begingroup$
Express $c_n$ from the recurrance formula: $c_n=g_n- frac1g_n+1$ taking the limits of both sides we have: $limlimits_nrightarrow infty(g_n- frac1g_n+1)=1tag1$ if $limlimits_nrightarrow inftyg_n=infty$ then (1) would not be valid. So the $limlimits_nrightarrow inftyg_n$ exists.
$endgroup$
– JV.Stalker
Mar 26 at 11:46
$begingroup$
Express $c_n$ from the recurrance formula: $c_n=g_n- frac1g_n+1$ taking the limits of both sides we have: $limlimits_nrightarrow infty(g_n- frac1g_n+1)=1tag1$ if $limlimits_nrightarrow inftyg_n=infty$ then (1) would not be valid. So the $limlimits_nrightarrow inftyg_n$ exists.
$endgroup$
– JV.Stalker
Mar 26 at 11:46
$begingroup$
Thanks for the comment. But doesn't this only prove that $lim_n to infty g_n$ does not equal infinity? Why does this prove that the limit $lim g_n$ exists?
$endgroup$
– Shibi Vasudevan
Mar 26 at 12:01
$begingroup$
Thanks for the comment. But doesn't this only prove that $lim_n to infty g_n$ does not equal infinity? Why does this prove that the limit $lim g_n$ exists?
$endgroup$
– Shibi Vasudevan
Mar 26 at 12:01
add a comment |
2 Answers
2
active
oldest
votes
$begingroup$
Lemma: If $u_n$ and $v_n$ are sequences of positive reals and $$lim_n → ∞ u_n = u > 0,quad varlimsup_n → ∞ v_n = v > 0,$$
then $varliminflimits_n → ∞ dfracu_nv_n = dfracuv$.
Proof: On the one hand, since $varlimsuplimits_n → ∞ v_n = v$, there exists a subsequence $v_n_k$ such that $limlimits_n → ∞ v_n_k = v$, then$$
varliminf_n → ∞ fracu_nv_n = lim_m → ∞ inf_n geqslant m fracu_nv_n leqslant lim_m → ∞ inf_k geqslant m fracu_n_kv_n_k = varliminf_k → ∞ fracu_n_kv_n_k = lim_k → ∞ fracu_n_kv_n_k = fracuv.
$$
On the other hand, for any $0 < ε < 1$, there exists $N geqslant 1$ such that $u_n > (1 - ε)u$ for $n geqslant N$. Thus for $m geqslant N$,$$
inf_n geqslant m fracu_nv_n geqslant inf_n geqslant m frac(1 - ε)uv_n = frac(1 - ε)usuplimits_n geqslant m v_n,
$$
which implies$$
varliminf_n → ∞ fracu_nv_n = lim_m → ∞ inf_n geqslant m fracu_nv_n geqslant (1 - ε)u lim_m → ∞ frac1suplimits_n geqslant m v_n = frac(1 - ε)ulimlimits_m → ∞ suplimits_n geqslant m v_n = (1 - ε) fracuv,
$$
and making $ε → 0^+$ yields $varliminflimits_n → ∞ dfracu_nv_n geqslant dfracuv$. Therefore, $varliminflimits_n → ∞ dfracu_nv_n = dfracuv$.
Now return to the question. Define $a_n = -g_n$, then $a_n + 1 = dfrac1a_n + c_n$ and $a_1 = c_1$. In fact, a more general proposition can be proved.
Proposition: Suppose that $b_n$ and $c_n$ are sequences of positive reals and$$
lim_n → ∞ b_n = b > 0,quad lim_n → ∞ c_n = c > 0.
$$
If $a_n$ satisfies $a_1 > 0$ and $a_n + 1 = dfracb_na_n + c_n$ for $n geqslant 1$, then $limlimits_n → ∞ a_n$ exists.
Proof: First, it is easy to see that $a_n > 0$ for each $n$. Since there exists $N geqslant 1$ such that$$
fracb2 < b_n < 2b, quad fracc2 < c_n < 2c, quad forall n > N
$$
then for $n > N$,$$
a_n < fracb_nc_n < frac4bc Longrightarrow a_n > fracb_ndfrac4bc + c_n > fracbdfrac8bc + 4c,
$$
which implies that$$
l := varliminf_n → ∞ a_n geqslant fracbdfrac8bc + 4c > 0, quad L := varlimsup_n → ∞ a_n leqslant frac4bc < +∞.
$$
Now, making $n → ∞$ in $a_n + 1 = dfracb_na_n + c_n$ yields$$
l = varliminf_n → ∞ a_n + 1 = fraclimlimits_n → ∞ b_nvarlimsuplimits_n → ∞ a_n + limlimits_n → ∞ c_n = fracbL + c,
$$
and analogously $L = dfracbl + c$, thus$$
l = fracbc + dfracbc + l Longrightarrow l^2 + cl - b = 0 Longrightarrow l = frac12 (-c + sqrtc^2 + 4b),
$$
and analogously$$
L^2 + cL - b = 0 Longrightarrow L = frac12 (-c + sqrtc^2 + 4b).
$$
Therefore, $l = L$, which implies that $limlimits_n → ∞ a_n$ exists and is finite.
$endgroup$
$begingroup$
Thanks for your answer. Could you briefly indicate why $l=b/(L+c)$. ( I see that something like this must be true, just that it would be nice to have a proof of this.)
$endgroup$
– Shibi Vasudevan
Mar 30 at 16:24
$begingroup$
OK, thanks for that.
$endgroup$
– Shibi Vasudevan
Mar 30 at 16:38
$begingroup$
@ShibiVasudevan I've added the lemma.
$endgroup$
– Saad
Mar 30 at 17:01
$begingroup$
Thanks for the Lemma.
$endgroup$
– Shibi Vasudevan
Mar 30 at 17:39
add a comment |
$begingroup$
The way I would do this is as follows: first, assume $c_n=1$, for all $n$; then it's easy to show that $g_2n$ decreases, $g_2n+1$ increases, $-frac12 geq g_2k geq g_infty geq g_2m+1 geq -1, |g_2n-g_2n+1| to 0$ by standard manipulations, so $g_n to g_infty$ in this case.
Then wlog assume $|c_n-1| leq frac1100$ say for all $n geq 1$ or start from a place where that happens and renumber if you wish, so in particular $c_n geq frac99100$, call $h_n$ the sequence from step above (where $c_n=1$) and note that $g_n, h_n <0, h_n leq -frac12, h_n to g_infty$ from the above. Note also that $|g_n+1| leq frac1c_n < 2$ from our assumptions.
Then using $c_n-g_n geq frac99100, 1-h_n geq frac32$ and $|g_n+1-h_n+1| = |frac(c_n-1)+(h_n-g_n)((c_n-g_n)(1-h_n)| leq fracc_n-1frac32frac99100$, it follows that $|g_n+1-h_n+1| leq frac34(|c_n-1|+|h_n-g_n|)$. Let $d_n=|h_n-g_n| leq |h_n|+|g_n| leq 3, q_n=|c_n-1|, d_n, q_n geq 0, q_n to 0 $ and $d_n+1 leq frac34(q_n+ d_n)$; then a standard argument presented below shows $d_n to 0$ so $g_n-h_n to 0$, so $g_n to g_infty$.
Noting that $d_n+k+1 leq (frac34)^kd_n+(sup_pgeq nq_p)Sigma(frac3 4)^m leq 3(frac34)^k + 4(sup_pgeq nq_p)$, for any $epsilon >0$, we pick first $n_0$, s.t. $(sup_pgeq n_0q_p) < fracepsilon8$ and then $k_0$ st. $3(frac34)^k_0 < fracepsilon2$ and then $d_m < epsilon, m>n_0+k_0$, so $d_n to 0$ indeed and we are done.
$endgroup$
$begingroup$
Thanks very much for your answer. It contains ideas and calculations that are quite useful to me. Thanks again.
$endgroup$
– Shibi Vasudevan
Mar 30 at 17:49
$begingroup$
You are welcome
$endgroup$
– Conrad
Mar 30 at 17:50
add a comment |
Your Answer
StackExchange.ready(function()
var channelOptions =
tags: "".split(" "),
id: "69"
;
initTagRenderer("".split(" "), "".split(" "), channelOptions);
StackExchange.using("externalEditor", function()
// Have to fire editor after snippets, if snippets enabled
if (StackExchange.settings.snippets.snippetsEnabled)
StackExchange.using("snippets", function()
createEditor();
);
else
createEditor();
);
function createEditor()
StackExchange.prepareEditor(
heartbeatType: 'answer',
autoActivateHeartbeat: false,
convertImagesToLinks: true,
noModals: true,
showLowRepImageUploadWarning: true,
reputationToPostImages: 10,
bindNavPrevention: true,
postfix: "",
imageUploader:
brandingHtml: "Powered by u003ca class="icon-imgur-white" href="https://imgur.com/"u003eu003c/au003e",
contentPolicyHtml: "User contributions licensed under u003ca href="https://creativecommons.org/licenses/by-sa/3.0/"u003ecc by-sa 3.0 with attribution requiredu003c/au003e u003ca href="https://stackoverflow.com/legal/content-policy"u003e(content policy)u003c/au003e",
allowUrls: true
,
noCode: true, onDemand: true,
discardSelector: ".discard-answer"
,immediatelyShowMarkdownHelp:true
);
);
Sign up or log in
StackExchange.ready(function ()
StackExchange.helpers.onClickDraftSave('#login-link');
);
Sign up using Google
Sign up using Facebook
Sign up using Email and Password
Post as a guest
Required, but never shown
StackExchange.ready(
function ()
StackExchange.openid.initPostLogin('.new-post-login', 'https%3a%2f%2fmath.stackexchange.com%2fquestions%2f3162938%2flimit-of-continued-fraction%23new-answer', 'question_page');
);
Post as a guest
Required, but never shown
2 Answers
2
active
oldest
votes
2 Answers
2
active
oldest
votes
active
oldest
votes
active
oldest
votes
$begingroup$
Lemma: If $u_n$ and $v_n$ are sequences of positive reals and $$lim_n → ∞ u_n = u > 0,quad varlimsup_n → ∞ v_n = v > 0,$$
then $varliminflimits_n → ∞ dfracu_nv_n = dfracuv$.
Proof: On the one hand, since $varlimsuplimits_n → ∞ v_n = v$, there exists a subsequence $v_n_k$ such that $limlimits_n → ∞ v_n_k = v$, then$$
varliminf_n → ∞ fracu_nv_n = lim_m → ∞ inf_n geqslant m fracu_nv_n leqslant lim_m → ∞ inf_k geqslant m fracu_n_kv_n_k = varliminf_k → ∞ fracu_n_kv_n_k = lim_k → ∞ fracu_n_kv_n_k = fracuv.
$$
On the other hand, for any $0 < ε < 1$, there exists $N geqslant 1$ such that $u_n > (1 - ε)u$ for $n geqslant N$. Thus for $m geqslant N$,$$
inf_n geqslant m fracu_nv_n geqslant inf_n geqslant m frac(1 - ε)uv_n = frac(1 - ε)usuplimits_n geqslant m v_n,
$$
which implies$$
varliminf_n → ∞ fracu_nv_n = lim_m → ∞ inf_n geqslant m fracu_nv_n geqslant (1 - ε)u lim_m → ∞ frac1suplimits_n geqslant m v_n = frac(1 - ε)ulimlimits_m → ∞ suplimits_n geqslant m v_n = (1 - ε) fracuv,
$$
and making $ε → 0^+$ yields $varliminflimits_n → ∞ dfracu_nv_n geqslant dfracuv$. Therefore, $varliminflimits_n → ∞ dfracu_nv_n = dfracuv$.
Now return to the question. Define $a_n = -g_n$, then $a_n + 1 = dfrac1a_n + c_n$ and $a_1 = c_1$. In fact, a more general proposition can be proved.
Proposition: Suppose that $b_n$ and $c_n$ are sequences of positive reals and$$
lim_n → ∞ b_n = b > 0,quad lim_n → ∞ c_n = c > 0.
$$
If $a_n$ satisfies $a_1 > 0$ and $a_n + 1 = dfracb_na_n + c_n$ for $n geqslant 1$, then $limlimits_n → ∞ a_n$ exists.
Proof: First, it is easy to see that $a_n > 0$ for each $n$. Since there exists $N geqslant 1$ such that$$
fracb2 < b_n < 2b, quad fracc2 < c_n < 2c, quad forall n > N
$$
then for $n > N$,$$
a_n < fracb_nc_n < frac4bc Longrightarrow a_n > fracb_ndfrac4bc + c_n > fracbdfrac8bc + 4c,
$$
which implies that$$
l := varliminf_n → ∞ a_n geqslant fracbdfrac8bc + 4c > 0, quad L := varlimsup_n → ∞ a_n leqslant frac4bc < +∞.
$$
Now, making $n → ∞$ in $a_n + 1 = dfracb_na_n + c_n$ yields$$
l = varliminf_n → ∞ a_n + 1 = fraclimlimits_n → ∞ b_nvarlimsuplimits_n → ∞ a_n + limlimits_n → ∞ c_n = fracbL + c,
$$
and analogously $L = dfracbl + c$, thus$$
l = fracbc + dfracbc + l Longrightarrow l^2 + cl - b = 0 Longrightarrow l = frac12 (-c + sqrtc^2 + 4b),
$$
and analogously$$
L^2 + cL - b = 0 Longrightarrow L = frac12 (-c + sqrtc^2 + 4b).
$$
Therefore, $l = L$, which implies that $limlimits_n → ∞ a_n$ exists and is finite.
$endgroup$
$begingroup$
Thanks for your answer. Could you briefly indicate why $l=b/(L+c)$. ( I see that something like this must be true, just that it would be nice to have a proof of this.)
$endgroup$
– Shibi Vasudevan
Mar 30 at 16:24
$begingroup$
OK, thanks for that.
$endgroup$
– Shibi Vasudevan
Mar 30 at 16:38
$begingroup$
@ShibiVasudevan I've added the lemma.
$endgroup$
– Saad
Mar 30 at 17:01
$begingroup$
Thanks for the Lemma.
$endgroup$
– Shibi Vasudevan
Mar 30 at 17:39
add a comment |
$begingroup$
Lemma: If $u_n$ and $v_n$ are sequences of positive reals and $$lim_n → ∞ u_n = u > 0,quad varlimsup_n → ∞ v_n = v > 0,$$
then $varliminflimits_n → ∞ dfracu_nv_n = dfracuv$.
Proof: On the one hand, since $varlimsuplimits_n → ∞ v_n = v$, there exists a subsequence $v_n_k$ such that $limlimits_n → ∞ v_n_k = v$, then$$
varliminf_n → ∞ fracu_nv_n = lim_m → ∞ inf_n geqslant m fracu_nv_n leqslant lim_m → ∞ inf_k geqslant m fracu_n_kv_n_k = varliminf_k → ∞ fracu_n_kv_n_k = lim_k → ∞ fracu_n_kv_n_k = fracuv.
$$
On the other hand, for any $0 < ε < 1$, there exists $N geqslant 1$ such that $u_n > (1 - ε)u$ for $n geqslant N$. Thus for $m geqslant N$,$$
inf_n geqslant m fracu_nv_n geqslant inf_n geqslant m frac(1 - ε)uv_n = frac(1 - ε)usuplimits_n geqslant m v_n,
$$
which implies$$
varliminf_n → ∞ fracu_nv_n = lim_m → ∞ inf_n geqslant m fracu_nv_n geqslant (1 - ε)u lim_m → ∞ frac1suplimits_n geqslant m v_n = frac(1 - ε)ulimlimits_m → ∞ suplimits_n geqslant m v_n = (1 - ε) fracuv,
$$
and making $ε → 0^+$ yields $varliminflimits_n → ∞ dfracu_nv_n geqslant dfracuv$. Therefore, $varliminflimits_n → ∞ dfracu_nv_n = dfracuv$.
Now return to the question. Define $a_n = -g_n$, then $a_n + 1 = dfrac1a_n + c_n$ and $a_1 = c_1$. In fact, a more general proposition can be proved.
Proposition: Suppose that $b_n$ and $c_n$ are sequences of positive reals and$$
lim_n → ∞ b_n = b > 0,quad lim_n → ∞ c_n = c > 0.
$$
If $a_n$ satisfies $a_1 > 0$ and $a_n + 1 = dfracb_na_n + c_n$ for $n geqslant 1$, then $limlimits_n → ∞ a_n$ exists.
Proof: First, it is easy to see that $a_n > 0$ for each $n$. Since there exists $N geqslant 1$ such that$$
fracb2 < b_n < 2b, quad fracc2 < c_n < 2c, quad forall n > N
$$
then for $n > N$,$$
a_n < fracb_nc_n < frac4bc Longrightarrow a_n > fracb_ndfrac4bc + c_n > fracbdfrac8bc + 4c,
$$
which implies that$$
l := varliminf_n → ∞ a_n geqslant fracbdfrac8bc + 4c > 0, quad L := varlimsup_n → ∞ a_n leqslant frac4bc < +∞.
$$
Now, making $n → ∞$ in $a_n + 1 = dfracb_na_n + c_n$ yields$$
l = varliminf_n → ∞ a_n + 1 = fraclimlimits_n → ∞ b_nvarlimsuplimits_n → ∞ a_n + limlimits_n → ∞ c_n = fracbL + c,
$$
and analogously $L = dfracbl + c$, thus$$
l = fracbc + dfracbc + l Longrightarrow l^2 + cl - b = 0 Longrightarrow l = frac12 (-c + sqrtc^2 + 4b),
$$
and analogously$$
L^2 + cL - b = 0 Longrightarrow L = frac12 (-c + sqrtc^2 + 4b).
$$
Therefore, $l = L$, which implies that $limlimits_n → ∞ a_n$ exists and is finite.
$endgroup$
$begingroup$
Thanks for your answer. Could you briefly indicate why $l=b/(L+c)$. ( I see that something like this must be true, just that it would be nice to have a proof of this.)
$endgroup$
– Shibi Vasudevan
Mar 30 at 16:24
$begingroup$
OK, thanks for that.
$endgroup$
– Shibi Vasudevan
Mar 30 at 16:38
$begingroup$
@ShibiVasudevan I've added the lemma.
$endgroup$
– Saad
Mar 30 at 17:01
$begingroup$
Thanks for the Lemma.
$endgroup$
– Shibi Vasudevan
Mar 30 at 17:39
add a comment |
$begingroup$
Lemma: If $u_n$ and $v_n$ are sequences of positive reals and $$lim_n → ∞ u_n = u > 0,quad varlimsup_n → ∞ v_n = v > 0,$$
then $varliminflimits_n → ∞ dfracu_nv_n = dfracuv$.
Proof: On the one hand, since $varlimsuplimits_n → ∞ v_n = v$, there exists a subsequence $v_n_k$ such that $limlimits_n → ∞ v_n_k = v$, then$$
varliminf_n → ∞ fracu_nv_n = lim_m → ∞ inf_n geqslant m fracu_nv_n leqslant lim_m → ∞ inf_k geqslant m fracu_n_kv_n_k = varliminf_k → ∞ fracu_n_kv_n_k = lim_k → ∞ fracu_n_kv_n_k = fracuv.
$$
On the other hand, for any $0 < ε < 1$, there exists $N geqslant 1$ such that $u_n > (1 - ε)u$ for $n geqslant N$. Thus for $m geqslant N$,$$
inf_n geqslant m fracu_nv_n geqslant inf_n geqslant m frac(1 - ε)uv_n = frac(1 - ε)usuplimits_n geqslant m v_n,
$$
which implies$$
varliminf_n → ∞ fracu_nv_n = lim_m → ∞ inf_n geqslant m fracu_nv_n geqslant (1 - ε)u lim_m → ∞ frac1suplimits_n geqslant m v_n = frac(1 - ε)ulimlimits_m → ∞ suplimits_n geqslant m v_n = (1 - ε) fracuv,
$$
and making $ε → 0^+$ yields $varliminflimits_n → ∞ dfracu_nv_n geqslant dfracuv$. Therefore, $varliminflimits_n → ∞ dfracu_nv_n = dfracuv$.
Now return to the question. Define $a_n = -g_n$, then $a_n + 1 = dfrac1a_n + c_n$ and $a_1 = c_1$. In fact, a more general proposition can be proved.
Proposition: Suppose that $b_n$ and $c_n$ are sequences of positive reals and$$
lim_n → ∞ b_n = b > 0,quad lim_n → ∞ c_n = c > 0.
$$
If $a_n$ satisfies $a_1 > 0$ and $a_n + 1 = dfracb_na_n + c_n$ for $n geqslant 1$, then $limlimits_n → ∞ a_n$ exists.
Proof: First, it is easy to see that $a_n > 0$ for each $n$. Since there exists $N geqslant 1$ such that$$
fracb2 < b_n < 2b, quad fracc2 < c_n < 2c, quad forall n > N
$$
then for $n > N$,$$
a_n < fracb_nc_n < frac4bc Longrightarrow a_n > fracb_ndfrac4bc + c_n > fracbdfrac8bc + 4c,
$$
which implies that$$
l := varliminf_n → ∞ a_n geqslant fracbdfrac8bc + 4c > 0, quad L := varlimsup_n → ∞ a_n leqslant frac4bc < +∞.
$$
Now, making $n → ∞$ in $a_n + 1 = dfracb_na_n + c_n$ yields$$
l = varliminf_n → ∞ a_n + 1 = fraclimlimits_n → ∞ b_nvarlimsuplimits_n → ∞ a_n + limlimits_n → ∞ c_n = fracbL + c,
$$
and analogously $L = dfracbl + c$, thus$$
l = fracbc + dfracbc + l Longrightarrow l^2 + cl - b = 0 Longrightarrow l = frac12 (-c + sqrtc^2 + 4b),
$$
and analogously$$
L^2 + cL - b = 0 Longrightarrow L = frac12 (-c + sqrtc^2 + 4b).
$$
Therefore, $l = L$, which implies that $limlimits_n → ∞ a_n$ exists and is finite.
$endgroup$
Lemma: If $u_n$ and $v_n$ are sequences of positive reals and $$lim_n → ∞ u_n = u > 0,quad varlimsup_n → ∞ v_n = v > 0,$$
then $varliminflimits_n → ∞ dfracu_nv_n = dfracuv$.
Proof: On the one hand, since $varlimsuplimits_n → ∞ v_n = v$, there exists a subsequence $v_n_k$ such that $limlimits_n → ∞ v_n_k = v$, then$$
varliminf_n → ∞ fracu_nv_n = lim_m → ∞ inf_n geqslant m fracu_nv_n leqslant lim_m → ∞ inf_k geqslant m fracu_n_kv_n_k = varliminf_k → ∞ fracu_n_kv_n_k = lim_k → ∞ fracu_n_kv_n_k = fracuv.
$$
On the other hand, for any $0 < ε < 1$, there exists $N geqslant 1$ such that $u_n > (1 - ε)u$ for $n geqslant N$. Thus for $m geqslant N$,$$
inf_n geqslant m fracu_nv_n geqslant inf_n geqslant m frac(1 - ε)uv_n = frac(1 - ε)usuplimits_n geqslant m v_n,
$$
which implies$$
varliminf_n → ∞ fracu_nv_n = lim_m → ∞ inf_n geqslant m fracu_nv_n geqslant (1 - ε)u lim_m → ∞ frac1suplimits_n geqslant m v_n = frac(1 - ε)ulimlimits_m → ∞ suplimits_n geqslant m v_n = (1 - ε) fracuv,
$$
and making $ε → 0^+$ yields $varliminflimits_n → ∞ dfracu_nv_n geqslant dfracuv$. Therefore, $varliminflimits_n → ∞ dfracu_nv_n = dfracuv$.
Now return to the question. Define $a_n = -g_n$, then $a_n + 1 = dfrac1a_n + c_n$ and $a_1 = c_1$. In fact, a more general proposition can be proved.
Proposition: Suppose that $b_n$ and $c_n$ are sequences of positive reals and$$
lim_n → ∞ b_n = b > 0,quad lim_n → ∞ c_n = c > 0.
$$
If $a_n$ satisfies $a_1 > 0$ and $a_n + 1 = dfracb_na_n + c_n$ for $n geqslant 1$, then $limlimits_n → ∞ a_n$ exists.
Proof: First, it is easy to see that $a_n > 0$ for each $n$. Since there exists $N geqslant 1$ such that$$
fracb2 < b_n < 2b, quad fracc2 < c_n < 2c, quad forall n > N
$$
then for $n > N$,$$
a_n < fracb_nc_n < frac4bc Longrightarrow a_n > fracb_ndfrac4bc + c_n > fracbdfrac8bc + 4c,
$$
which implies that$$
l := varliminf_n → ∞ a_n geqslant fracbdfrac8bc + 4c > 0, quad L := varlimsup_n → ∞ a_n leqslant frac4bc < +∞.
$$
Now, making $n → ∞$ in $a_n + 1 = dfracb_na_n + c_n$ yields$$
l = varliminf_n → ∞ a_n + 1 = fraclimlimits_n → ∞ b_nvarlimsuplimits_n → ∞ a_n + limlimits_n → ∞ c_n = fracbL + c,
$$
and analogously $L = dfracbl + c$, thus$$
l = fracbc + dfracbc + l Longrightarrow l^2 + cl - b = 0 Longrightarrow l = frac12 (-c + sqrtc^2 + 4b),
$$
and analogously$$
L^2 + cL - b = 0 Longrightarrow L = frac12 (-c + sqrtc^2 + 4b).
$$
Therefore, $l = L$, which implies that $limlimits_n → ∞ a_n$ exists and is finite.
edited Mar 30 at 17:00
answered Mar 29 at 16:32
SaadSaad
20.6k92452
20.6k92452
$begingroup$
Thanks for your answer. Could you briefly indicate why $l=b/(L+c)$. ( I see that something like this must be true, just that it would be nice to have a proof of this.)
$endgroup$
– Shibi Vasudevan
Mar 30 at 16:24
$begingroup$
OK, thanks for that.
$endgroup$
– Shibi Vasudevan
Mar 30 at 16:38
$begingroup$
@ShibiVasudevan I've added the lemma.
$endgroup$
– Saad
Mar 30 at 17:01
$begingroup$
Thanks for the Lemma.
$endgroup$
– Shibi Vasudevan
Mar 30 at 17:39
add a comment |
$begingroup$
Thanks for your answer. Could you briefly indicate why $l=b/(L+c)$. ( I see that something like this must be true, just that it would be nice to have a proof of this.)
$endgroup$
– Shibi Vasudevan
Mar 30 at 16:24
$begingroup$
OK, thanks for that.
$endgroup$
– Shibi Vasudevan
Mar 30 at 16:38
$begingroup$
@ShibiVasudevan I've added the lemma.
$endgroup$
– Saad
Mar 30 at 17:01
$begingroup$
Thanks for the Lemma.
$endgroup$
– Shibi Vasudevan
Mar 30 at 17:39
$begingroup$
Thanks for your answer. Could you briefly indicate why $l=b/(L+c)$. ( I see that something like this must be true, just that it would be nice to have a proof of this.)
$endgroup$
– Shibi Vasudevan
Mar 30 at 16:24
$begingroup$
Thanks for your answer. Could you briefly indicate why $l=b/(L+c)$. ( I see that something like this must be true, just that it would be nice to have a proof of this.)
$endgroup$
– Shibi Vasudevan
Mar 30 at 16:24
$begingroup$
OK, thanks for that.
$endgroup$
– Shibi Vasudevan
Mar 30 at 16:38
$begingroup$
OK, thanks for that.
$endgroup$
– Shibi Vasudevan
Mar 30 at 16:38
$begingroup$
@ShibiVasudevan I've added the lemma.
$endgroup$
– Saad
Mar 30 at 17:01
$begingroup$
@ShibiVasudevan I've added the lemma.
$endgroup$
– Saad
Mar 30 at 17:01
$begingroup$
Thanks for the Lemma.
$endgroup$
– Shibi Vasudevan
Mar 30 at 17:39
$begingroup$
Thanks for the Lemma.
$endgroup$
– Shibi Vasudevan
Mar 30 at 17:39
add a comment |
$begingroup$
The way I would do this is as follows: first, assume $c_n=1$, for all $n$; then it's easy to show that $g_2n$ decreases, $g_2n+1$ increases, $-frac12 geq g_2k geq g_infty geq g_2m+1 geq -1, |g_2n-g_2n+1| to 0$ by standard manipulations, so $g_n to g_infty$ in this case.
Then wlog assume $|c_n-1| leq frac1100$ say for all $n geq 1$ or start from a place where that happens and renumber if you wish, so in particular $c_n geq frac99100$, call $h_n$ the sequence from step above (where $c_n=1$) and note that $g_n, h_n <0, h_n leq -frac12, h_n to g_infty$ from the above. Note also that $|g_n+1| leq frac1c_n < 2$ from our assumptions.
Then using $c_n-g_n geq frac99100, 1-h_n geq frac32$ and $|g_n+1-h_n+1| = |frac(c_n-1)+(h_n-g_n)((c_n-g_n)(1-h_n)| leq fracc_n-1frac32frac99100$, it follows that $|g_n+1-h_n+1| leq frac34(|c_n-1|+|h_n-g_n|)$. Let $d_n=|h_n-g_n| leq |h_n|+|g_n| leq 3, q_n=|c_n-1|, d_n, q_n geq 0, q_n to 0 $ and $d_n+1 leq frac34(q_n+ d_n)$; then a standard argument presented below shows $d_n to 0$ so $g_n-h_n to 0$, so $g_n to g_infty$.
Noting that $d_n+k+1 leq (frac34)^kd_n+(sup_pgeq nq_p)Sigma(frac3 4)^m leq 3(frac34)^k + 4(sup_pgeq nq_p)$, for any $epsilon >0$, we pick first $n_0$, s.t. $(sup_pgeq n_0q_p) < fracepsilon8$ and then $k_0$ st. $3(frac34)^k_0 < fracepsilon2$ and then $d_m < epsilon, m>n_0+k_0$, so $d_n to 0$ indeed and we are done.
$endgroup$
$begingroup$
Thanks very much for your answer. It contains ideas and calculations that are quite useful to me. Thanks again.
$endgroup$
– Shibi Vasudevan
Mar 30 at 17:49
$begingroup$
You are welcome
$endgroup$
– Conrad
Mar 30 at 17:50
add a comment |
$begingroup$
The way I would do this is as follows: first, assume $c_n=1$, for all $n$; then it's easy to show that $g_2n$ decreases, $g_2n+1$ increases, $-frac12 geq g_2k geq g_infty geq g_2m+1 geq -1, |g_2n-g_2n+1| to 0$ by standard manipulations, so $g_n to g_infty$ in this case.
Then wlog assume $|c_n-1| leq frac1100$ say for all $n geq 1$ or start from a place where that happens and renumber if you wish, so in particular $c_n geq frac99100$, call $h_n$ the sequence from step above (where $c_n=1$) and note that $g_n, h_n <0, h_n leq -frac12, h_n to g_infty$ from the above. Note also that $|g_n+1| leq frac1c_n < 2$ from our assumptions.
Then using $c_n-g_n geq frac99100, 1-h_n geq frac32$ and $|g_n+1-h_n+1| = |frac(c_n-1)+(h_n-g_n)((c_n-g_n)(1-h_n)| leq fracc_n-1frac32frac99100$, it follows that $|g_n+1-h_n+1| leq frac34(|c_n-1|+|h_n-g_n|)$. Let $d_n=|h_n-g_n| leq |h_n|+|g_n| leq 3, q_n=|c_n-1|, d_n, q_n geq 0, q_n to 0 $ and $d_n+1 leq frac34(q_n+ d_n)$; then a standard argument presented below shows $d_n to 0$ so $g_n-h_n to 0$, so $g_n to g_infty$.
Noting that $d_n+k+1 leq (frac34)^kd_n+(sup_pgeq nq_p)Sigma(frac3 4)^m leq 3(frac34)^k + 4(sup_pgeq nq_p)$, for any $epsilon >0$, we pick first $n_0$, s.t. $(sup_pgeq n_0q_p) < fracepsilon8$ and then $k_0$ st. $3(frac34)^k_0 < fracepsilon2$ and then $d_m < epsilon, m>n_0+k_0$, so $d_n to 0$ indeed and we are done.
$endgroup$
$begingroup$
Thanks very much for your answer. It contains ideas and calculations that are quite useful to me. Thanks again.
$endgroup$
– Shibi Vasudevan
Mar 30 at 17:49
$begingroup$
You are welcome
$endgroup$
– Conrad
Mar 30 at 17:50
add a comment |
$begingroup$
The way I would do this is as follows: first, assume $c_n=1$, for all $n$; then it's easy to show that $g_2n$ decreases, $g_2n+1$ increases, $-frac12 geq g_2k geq g_infty geq g_2m+1 geq -1, |g_2n-g_2n+1| to 0$ by standard manipulations, so $g_n to g_infty$ in this case.
Then wlog assume $|c_n-1| leq frac1100$ say for all $n geq 1$ or start from a place where that happens and renumber if you wish, so in particular $c_n geq frac99100$, call $h_n$ the sequence from step above (where $c_n=1$) and note that $g_n, h_n <0, h_n leq -frac12, h_n to g_infty$ from the above. Note also that $|g_n+1| leq frac1c_n < 2$ from our assumptions.
Then using $c_n-g_n geq frac99100, 1-h_n geq frac32$ and $|g_n+1-h_n+1| = |frac(c_n-1)+(h_n-g_n)((c_n-g_n)(1-h_n)| leq fracc_n-1frac32frac99100$, it follows that $|g_n+1-h_n+1| leq frac34(|c_n-1|+|h_n-g_n|)$. Let $d_n=|h_n-g_n| leq |h_n|+|g_n| leq 3, q_n=|c_n-1|, d_n, q_n geq 0, q_n to 0 $ and $d_n+1 leq frac34(q_n+ d_n)$; then a standard argument presented below shows $d_n to 0$ so $g_n-h_n to 0$, so $g_n to g_infty$.
Noting that $d_n+k+1 leq (frac34)^kd_n+(sup_pgeq nq_p)Sigma(frac3 4)^m leq 3(frac34)^k + 4(sup_pgeq nq_p)$, for any $epsilon >0$, we pick first $n_0$, s.t. $(sup_pgeq n_0q_p) < fracepsilon8$ and then $k_0$ st. $3(frac34)^k_0 < fracepsilon2$ and then $d_m < epsilon, m>n_0+k_0$, so $d_n to 0$ indeed and we are done.
$endgroup$
The way I would do this is as follows: first, assume $c_n=1$, for all $n$; then it's easy to show that $g_2n$ decreases, $g_2n+1$ increases, $-frac12 geq g_2k geq g_infty geq g_2m+1 geq -1, |g_2n-g_2n+1| to 0$ by standard manipulations, so $g_n to g_infty$ in this case.
Then wlog assume $|c_n-1| leq frac1100$ say for all $n geq 1$ or start from a place where that happens and renumber if you wish, so in particular $c_n geq frac99100$, call $h_n$ the sequence from step above (where $c_n=1$) and note that $g_n, h_n <0, h_n leq -frac12, h_n to g_infty$ from the above. Note also that $|g_n+1| leq frac1c_n < 2$ from our assumptions.
Then using $c_n-g_n geq frac99100, 1-h_n geq frac32$ and $|g_n+1-h_n+1| = |frac(c_n-1)+(h_n-g_n)((c_n-g_n)(1-h_n)| leq fracc_n-1frac32frac99100$, it follows that $|g_n+1-h_n+1| leq frac34(|c_n-1|+|h_n-g_n|)$. Let $d_n=|h_n-g_n| leq |h_n|+|g_n| leq 3, q_n=|c_n-1|, d_n, q_n geq 0, q_n to 0 $ and $d_n+1 leq frac34(q_n+ d_n)$; then a standard argument presented below shows $d_n to 0$ so $g_n-h_n to 0$, so $g_n to g_infty$.
Noting that $d_n+k+1 leq (frac34)^kd_n+(sup_pgeq nq_p)Sigma(frac3 4)^m leq 3(frac34)^k + 4(sup_pgeq nq_p)$, for any $epsilon >0$, we pick first $n_0$, s.t. $(sup_pgeq n_0q_p) < fracepsilon8$ and then $k_0$ st. $3(frac34)^k_0 < fracepsilon2$ and then $d_m < epsilon, m>n_0+k_0$, so $d_n to 0$ indeed and we are done.
answered Mar 29 at 13:51
ConradConrad
1,51745
1,51745
$begingroup$
Thanks very much for your answer. It contains ideas and calculations that are quite useful to me. Thanks again.
$endgroup$
– Shibi Vasudevan
Mar 30 at 17:49
$begingroup$
You are welcome
$endgroup$
– Conrad
Mar 30 at 17:50
add a comment |
$begingroup$
Thanks very much for your answer. It contains ideas and calculations that are quite useful to me. Thanks again.
$endgroup$
– Shibi Vasudevan
Mar 30 at 17:49
$begingroup$
You are welcome
$endgroup$
– Conrad
Mar 30 at 17:50
$begingroup$
Thanks very much for your answer. It contains ideas and calculations that are quite useful to me. Thanks again.
$endgroup$
– Shibi Vasudevan
Mar 30 at 17:49
$begingroup$
Thanks very much for your answer. It contains ideas and calculations that are quite useful to me. Thanks again.
$endgroup$
– Shibi Vasudevan
Mar 30 at 17:49
$begingroup$
You are welcome
$endgroup$
– Conrad
Mar 30 at 17:50
$begingroup$
You are welcome
$endgroup$
– Conrad
Mar 30 at 17:50
add a comment |
Thanks for contributing an answer to Mathematics Stack Exchange!
- Please be sure to answer the question. Provide details and share your research!
But avoid …
- Asking for help, clarification, or responding to other answers.
- Making statements based on opinion; back them up with references or personal experience.
Use MathJax to format equations. MathJax reference.
To learn more, see our tips on writing great answers.
Sign up or log in
StackExchange.ready(function ()
StackExchange.helpers.onClickDraftSave('#login-link');
);
Sign up using Google
Sign up using Facebook
Sign up using Email and Password
Post as a guest
Required, but never shown
StackExchange.ready(
function ()
StackExchange.openid.initPostLogin('.new-post-login', 'https%3a%2f%2fmath.stackexchange.com%2fquestions%2f3162938%2flimit-of-continued-fraction%23new-answer', 'question_page');
);
Post as a guest
Required, but never shown
Sign up or log in
StackExchange.ready(function ()
StackExchange.helpers.onClickDraftSave('#login-link');
);
Sign up using Google
Sign up using Facebook
Sign up using Email and Password
Post as a guest
Required, but never shown
Sign up or log in
StackExchange.ready(function ()
StackExchange.helpers.onClickDraftSave('#login-link');
);
Sign up using Google
Sign up using Facebook
Sign up using Email and Password
Post as a guest
Required, but never shown
Sign up or log in
StackExchange.ready(function ()
StackExchange.helpers.onClickDraftSave('#login-link');
);
Sign up using Google
Sign up using Facebook
Sign up using Email and Password
Sign up using Google
Sign up using Facebook
Sign up using Email and Password
Post as a guest
Required, but never shown
Required, but never shown
Required, but never shown
Required, but never shown
Required, but never shown
Required, but never shown
Required, but never shown
Required, but never shown
Required, but never shown
$begingroup$
Express $c_n$ from the recurrance formula: $c_n=g_n- frac1g_n+1$ taking the limits of both sides we have: $limlimits_nrightarrow infty(g_n- frac1g_n+1)=1tag1$ if $limlimits_nrightarrow inftyg_n=infty$ then (1) would not be valid. So the $limlimits_nrightarrow inftyg_n$ exists.
$endgroup$
– JV.Stalker
Mar 26 at 11:46
$begingroup$
Thanks for the comment. But doesn't this only prove that $lim_n to infty g_n$ does not equal infinity? Why does this prove that the limit $lim g_n$ exists?
$endgroup$
– Shibi Vasudevan
Mar 26 at 12:01