How to test for convergence for the harmonic series with irregular (binomial) sign changes? [closed]

The name of the pictureThe name of the pictureThe name of the pictureClash Royale CLAN TAG#URR8PPP











up vote
1
down vote

favorite
1












How would you test for convergence with a series such as this using the alternating series test?



$1-frac12-frac13+frac14+frac15+frac16-frac17...-frac110+frac111...+frac115-....$



I have attempted to find a summation equation for this by grouping in different ways in order to use the alternating series test but I can't find a way to group these in a logical way. Is there a grouping that would allow for analysis of this series using the alternating series test?







share|cite|improve this question













closed as unclear what you're asking by Siong Thye Goh, Leucippus, Claude Leibovici, user91500, Henrik Sep 15 '17 at 11:39


Please clarify your specific problem or add additional details to highlight exactly what you need. As it's currently written, it’s hard to tell exactly what you're asking. See the How to Ask page for help clarifying this question. If this question can be reworded to fit the rules in the help center, please edit the question.














  • It appears that the sign changes are happening logically. The series alternates signs, but extends the length by 1 every time the sign changes. 1 positive, 2 negative, 3 positive, 4 negative, 5 positive, etc.
    – PhilTheLawyer
    Sep 15 '17 at 2:19






  • 1




    I would suggest grouping consecutive terms with the same sign, then attempting to apply the alternating series test.
    – Xander Henderson
    Sep 15 '17 at 2:30










  • For the alternating series test you need to look at the limit as n goes to infinity. The part I'm struggling with is a way to represent this series using n. I can see that each term decreases but the limit is where I'm getting caught up.
    – Ian B.
    Sep 15 '17 at 2:32










  • You didn't define a series but only its first few terms... And partial summation shows $sum_n=1^infty frac(-1)^a_nn$ converges whenever $lim_N to infty N^-csum_n=1^N (-1)^a_n = 0$ for some $c < 1$.
    – reuns
    Sep 15 '17 at 2:42











  • @IanBaughman Have you tried grouping terms as I suggested? Can you see the pattern?
    – Xander Henderson
    Sep 15 '17 at 2:54














up vote
1
down vote

favorite
1












How would you test for convergence with a series such as this using the alternating series test?



$1-frac12-frac13+frac14+frac15+frac16-frac17...-frac110+frac111...+frac115-....$



I have attempted to find a summation equation for this by grouping in different ways in order to use the alternating series test but I can't find a way to group these in a logical way. Is there a grouping that would allow for analysis of this series using the alternating series test?







share|cite|improve this question













closed as unclear what you're asking by Siong Thye Goh, Leucippus, Claude Leibovici, user91500, Henrik Sep 15 '17 at 11:39


Please clarify your specific problem or add additional details to highlight exactly what you need. As it's currently written, it’s hard to tell exactly what you're asking. See the How to Ask page for help clarifying this question. If this question can be reworded to fit the rules in the help center, please edit the question.














  • It appears that the sign changes are happening logically. The series alternates signs, but extends the length by 1 every time the sign changes. 1 positive, 2 negative, 3 positive, 4 negative, 5 positive, etc.
    – PhilTheLawyer
    Sep 15 '17 at 2:19






  • 1




    I would suggest grouping consecutive terms with the same sign, then attempting to apply the alternating series test.
    – Xander Henderson
    Sep 15 '17 at 2:30










  • For the alternating series test you need to look at the limit as n goes to infinity. The part I'm struggling with is a way to represent this series using n. I can see that each term decreases but the limit is where I'm getting caught up.
    – Ian B.
    Sep 15 '17 at 2:32










  • You didn't define a series but only its first few terms... And partial summation shows $sum_n=1^infty frac(-1)^a_nn$ converges whenever $lim_N to infty N^-csum_n=1^N (-1)^a_n = 0$ for some $c < 1$.
    – reuns
    Sep 15 '17 at 2:42











  • @IanBaughman Have you tried grouping terms as I suggested? Can you see the pattern?
    – Xander Henderson
    Sep 15 '17 at 2:54












up vote
1
down vote

favorite
1









up vote
1
down vote

favorite
1






1





How would you test for convergence with a series such as this using the alternating series test?



$1-frac12-frac13+frac14+frac15+frac16-frac17...-frac110+frac111...+frac115-....$



I have attempted to find a summation equation for this by grouping in different ways in order to use the alternating series test but I can't find a way to group these in a logical way. Is there a grouping that would allow for analysis of this series using the alternating series test?







share|cite|improve this question













How would you test for convergence with a series such as this using the alternating series test?



$1-frac12-frac13+frac14+frac15+frac16-frac17...-frac110+frac111...+frac115-....$



I have attempted to find a summation equation for this by grouping in different ways in order to use the alternating series test but I can't find a way to group these in a logical way. Is there a grouping that would allow for analysis of this series using the alternating series test?









share|cite|improve this question












share|cite|improve this question




share|cite|improve this question








edited Sep 23 '17 at 3:11









Parcly Taxel

33.6k136588




33.6k136588









asked Sep 15 '17 at 2:15









Ian B.

828




828




closed as unclear what you're asking by Siong Thye Goh, Leucippus, Claude Leibovici, user91500, Henrik Sep 15 '17 at 11:39


Please clarify your specific problem or add additional details to highlight exactly what you need. As it's currently written, it’s hard to tell exactly what you're asking. See the How to Ask page for help clarifying this question. If this question can be reworded to fit the rules in the help center, please edit the question.






closed as unclear what you're asking by Siong Thye Goh, Leucippus, Claude Leibovici, user91500, Henrik Sep 15 '17 at 11:39


Please clarify your specific problem or add additional details to highlight exactly what you need. As it's currently written, it’s hard to tell exactly what you're asking. See the How to Ask page for help clarifying this question. If this question can be reworded to fit the rules in the help center, please edit the question.













  • It appears that the sign changes are happening logically. The series alternates signs, but extends the length by 1 every time the sign changes. 1 positive, 2 negative, 3 positive, 4 negative, 5 positive, etc.
    – PhilTheLawyer
    Sep 15 '17 at 2:19






  • 1




    I would suggest grouping consecutive terms with the same sign, then attempting to apply the alternating series test.
    – Xander Henderson
    Sep 15 '17 at 2:30










  • For the alternating series test you need to look at the limit as n goes to infinity. The part I'm struggling with is a way to represent this series using n. I can see that each term decreases but the limit is where I'm getting caught up.
    – Ian B.
    Sep 15 '17 at 2:32










  • You didn't define a series but only its first few terms... And partial summation shows $sum_n=1^infty frac(-1)^a_nn$ converges whenever $lim_N to infty N^-csum_n=1^N (-1)^a_n = 0$ for some $c < 1$.
    – reuns
    Sep 15 '17 at 2:42











  • @IanBaughman Have you tried grouping terms as I suggested? Can you see the pattern?
    – Xander Henderson
    Sep 15 '17 at 2:54
















  • It appears that the sign changes are happening logically. The series alternates signs, but extends the length by 1 every time the sign changes. 1 positive, 2 negative, 3 positive, 4 negative, 5 positive, etc.
    – PhilTheLawyer
    Sep 15 '17 at 2:19






  • 1




    I would suggest grouping consecutive terms with the same sign, then attempting to apply the alternating series test.
    – Xander Henderson
    Sep 15 '17 at 2:30










  • For the alternating series test you need to look at the limit as n goes to infinity. The part I'm struggling with is a way to represent this series using n. I can see that each term decreases but the limit is where I'm getting caught up.
    – Ian B.
    Sep 15 '17 at 2:32










  • You didn't define a series but only its first few terms... And partial summation shows $sum_n=1^infty frac(-1)^a_nn$ converges whenever $lim_N to infty N^-csum_n=1^N (-1)^a_n = 0$ for some $c < 1$.
    – reuns
    Sep 15 '17 at 2:42











  • @IanBaughman Have you tried grouping terms as I suggested? Can you see the pattern?
    – Xander Henderson
    Sep 15 '17 at 2:54















It appears that the sign changes are happening logically. The series alternates signs, but extends the length by 1 every time the sign changes. 1 positive, 2 negative, 3 positive, 4 negative, 5 positive, etc.
– PhilTheLawyer
Sep 15 '17 at 2:19




It appears that the sign changes are happening logically. The series alternates signs, but extends the length by 1 every time the sign changes. 1 positive, 2 negative, 3 positive, 4 negative, 5 positive, etc.
– PhilTheLawyer
Sep 15 '17 at 2:19




1




1




I would suggest grouping consecutive terms with the same sign, then attempting to apply the alternating series test.
– Xander Henderson
Sep 15 '17 at 2:30




I would suggest grouping consecutive terms with the same sign, then attempting to apply the alternating series test.
– Xander Henderson
Sep 15 '17 at 2:30












For the alternating series test you need to look at the limit as n goes to infinity. The part I'm struggling with is a way to represent this series using n. I can see that each term decreases but the limit is where I'm getting caught up.
– Ian B.
Sep 15 '17 at 2:32




For the alternating series test you need to look at the limit as n goes to infinity. The part I'm struggling with is a way to represent this series using n. I can see that each term decreases but the limit is where I'm getting caught up.
– Ian B.
Sep 15 '17 at 2:32












You didn't define a series but only its first few terms... And partial summation shows $sum_n=1^infty frac(-1)^a_nn$ converges whenever $lim_N to infty N^-csum_n=1^N (-1)^a_n = 0$ for some $c < 1$.
– reuns
Sep 15 '17 at 2:42





You didn't define a series but only its first few terms... And partial summation shows $sum_n=1^infty frac(-1)^a_nn$ converges whenever $lim_N to infty N^-csum_n=1^N (-1)^a_n = 0$ for some $c < 1$.
– reuns
Sep 15 '17 at 2:42













@IanBaughman Have you tried grouping terms as I suggested? Can you see the pattern?
– Xander Henderson
Sep 15 '17 at 2:54




@IanBaughman Have you tried grouping terms as I suggested? Can you see the pattern?
– Xander Henderson
Sep 15 '17 at 2:54










2 Answers
2






active

oldest

votes

















up vote
1
down vote













I hope I'm interpreting this correctly: you have one $+$ term, then two $-$, three $+$, four $-$, etc. The start and end of a run of $2k+1$ $+$ terms are $1/(2 k^2 + k + 1)$ and $1/(2 k^2 + 3 k + 1)$ and the start and end of the next run of $2k+2$ $-1$ terms are $-1/(2 k^2 + 3 k + 2)$ and $-1/(2 k^2 + 5 k + 3$.
Thus the sum of these two runs is at least



$$ dfrac2k+12 k^2 + 3k + 1 - dfrac2k+22k^2+3k+2 = frac-k(2 k^2 + 3k + 1)(2k^2+3k+2)$$
and at most



$$ dfrac2k+12 k^2 +k + 1 - dfrac2k+22k^2+5k+3 = frac-k(2 k^2 + 3k + 1)(2k^2+3k+2) = frac6k+1(2k+3)(2k^2+k+1)$$



Thus in any case it is bounded in absolute value by $C/k^2$ for some $C$, and thus the series converges.






share|cite|improve this answer





















  • I get with pari/gp using the shortcut-definition for the binomial-expression x.b = 1+binomial(1+x,2) (so for instance 4.b = 1+binomial(5,2)=11) and using the psi()-function for partial-sums of harmonic numbers for some $N$ for instance N=5;psi(0.b)+(-1)^N*psi(N.b)-2*sum(k=0,N,(-1)^k*psi(k.b)) an alternating sequence of partial sums with decreasing terms. Euler-summation or using sumalt() gives a value of about $ 0.517100379042$
    – Gottfried Helms
    Sep 18 '17 at 10:04

















up vote
0
down vote













Define $$a_n = displaystylesum_k=fracn(n-1)2 + 1^fracn(n+1)2 frac 1k$$.



So for example, $a_1 = 1$, $a_2 = frac 12 + frac 13$, $a_3= frac 14 + frac 15 + frac 16$, etc.



It's easy to check that series you have written above is $sum_i=1^infty (-1)^i a_i$. It's easy to see that $a_i to 0$,since each fraction in the summation is less than $frac2n(n-1)+ 2$, and there are $n$ of them, so $a_i leq frac2nn(n-1) + 2$, which decreases to zero as $n to infty$.



You can also show that $a_n$ are decreasing as follows :
$a_n - a_n+1 = displaystylesum_k=fracn(n-1)2 + 1^fracn(n+1)2 frac 1k - displaystylesum_k=fracn(n+1)2 + 1^frac(n+2)(n+1)2 frac 1k$



What we do in this sum, is take terms whose difference is exactly $n$. More specifically, I claim that the above sum is the following:
$$
a_n - a_n+1 = frac-2(n+1)(n+2) + displaystylesum_k=fracn(n-1)2 + 1^fracn(n+1)2 left(frac 1k - frac1k+nright)
$$



As if to give an example, $a_3 - a_4 = frac 14 + frac 15 + frac 16 - frac 17 - frac 18 - frac 19 - frac 110 \ = frac 14 - frac 17 + frac 15 - frac 18 + frac 16 - frac 19 - frac 110$. Now you see why this sum holds.



Now, simplify the sums inside :
$$
a_n - a_n+1 = frac-2(n+1)(n+2) + displaystylesum_k=fracn(n-1)2 + 1^fracn(n+1)2 left(fracnk(k+n)right)
$$
Each term on the right hand side is greater than the term with $k = fracn(n+1)2$, and there are $n$ such terms, so:
$$
a_n - a_n+1 geq frac-2(n+1)(n+2) + nleft(fracnfracn(n+1)2left(fracn(n+1)2+nright)right)
$$
which after plenty of simplification becomes:
$$
a_n - a_n+1 geq frac-2(n+1)(n+2) + frac4(n+1)(n+3)
$$
which can be written as:
$$
a_n - a_n+1 geq 4 left(frac1(n+1)(n+3) - frac1(n+1)(2n+2)right)
$$
and the right hand side is positive when $n+3 leq 2n+2 $ i.e. when $n geq 1$.
For example, $a_3 - a_4 $ is predicted to be greater than $frac124$ by this estimate, and it's actual value is $frac3472520$, which is clearly greater than $frac 124$.



So, it follows that $a_n$ is a decreasing sequence which converges to $0$. Now, the alternate series test is applicable, I would think, and this shows that the given series converges conditionally.



Please point out any mistake if found.
$()$






share|cite|improve this answer





















  • Thank you for the acceptance. I would like you to ask somebody to actually compute this sum. I tried doing so on Wolfram Alpha, but it exceeded standard time, so I ran into trouble. That value will settle this question once and for all.
    – Ð°ÑÑ‚он вілла олоф мэллбэрг
    Sep 15 '17 at 3:28

















2 Answers
2






active

oldest

votes








2 Answers
2






active

oldest

votes









active

oldest

votes






active

oldest

votes








up vote
1
down vote













I hope I'm interpreting this correctly: you have one $+$ term, then two $-$, three $+$, four $-$, etc. The start and end of a run of $2k+1$ $+$ terms are $1/(2 k^2 + k + 1)$ and $1/(2 k^2 + 3 k + 1)$ and the start and end of the next run of $2k+2$ $-1$ terms are $-1/(2 k^2 + 3 k + 2)$ and $-1/(2 k^2 + 5 k + 3$.
Thus the sum of these two runs is at least



$$ dfrac2k+12 k^2 + 3k + 1 - dfrac2k+22k^2+3k+2 = frac-k(2 k^2 + 3k + 1)(2k^2+3k+2)$$
and at most



$$ dfrac2k+12 k^2 +k + 1 - dfrac2k+22k^2+5k+3 = frac-k(2 k^2 + 3k + 1)(2k^2+3k+2) = frac6k+1(2k+3)(2k^2+k+1)$$



Thus in any case it is bounded in absolute value by $C/k^2$ for some $C$, and thus the series converges.






share|cite|improve this answer





















  • I get with pari/gp using the shortcut-definition for the binomial-expression x.b = 1+binomial(1+x,2) (so for instance 4.b = 1+binomial(5,2)=11) and using the psi()-function for partial-sums of harmonic numbers for some $N$ for instance N=5;psi(0.b)+(-1)^N*psi(N.b)-2*sum(k=0,N,(-1)^k*psi(k.b)) an alternating sequence of partial sums with decreasing terms. Euler-summation or using sumalt() gives a value of about $ 0.517100379042$
    – Gottfried Helms
    Sep 18 '17 at 10:04














up vote
1
down vote













I hope I'm interpreting this correctly: you have one $+$ term, then two $-$, three $+$, four $-$, etc. The start and end of a run of $2k+1$ $+$ terms are $1/(2 k^2 + k + 1)$ and $1/(2 k^2 + 3 k + 1)$ and the start and end of the next run of $2k+2$ $-1$ terms are $-1/(2 k^2 + 3 k + 2)$ and $-1/(2 k^2 + 5 k + 3$.
Thus the sum of these two runs is at least



$$ dfrac2k+12 k^2 + 3k + 1 - dfrac2k+22k^2+3k+2 = frac-k(2 k^2 + 3k + 1)(2k^2+3k+2)$$
and at most



$$ dfrac2k+12 k^2 +k + 1 - dfrac2k+22k^2+5k+3 = frac-k(2 k^2 + 3k + 1)(2k^2+3k+2) = frac6k+1(2k+3)(2k^2+k+1)$$



Thus in any case it is bounded in absolute value by $C/k^2$ for some $C$, and thus the series converges.






share|cite|improve this answer





















  • I get with pari/gp using the shortcut-definition for the binomial-expression x.b = 1+binomial(1+x,2) (so for instance 4.b = 1+binomial(5,2)=11) and using the psi()-function for partial-sums of harmonic numbers for some $N$ for instance N=5;psi(0.b)+(-1)^N*psi(N.b)-2*sum(k=0,N,(-1)^k*psi(k.b)) an alternating sequence of partial sums with decreasing terms. Euler-summation or using sumalt() gives a value of about $ 0.517100379042$
    – Gottfried Helms
    Sep 18 '17 at 10:04












up vote
1
down vote










up vote
1
down vote









I hope I'm interpreting this correctly: you have one $+$ term, then two $-$, three $+$, four $-$, etc. The start and end of a run of $2k+1$ $+$ terms are $1/(2 k^2 + k + 1)$ and $1/(2 k^2 + 3 k + 1)$ and the start and end of the next run of $2k+2$ $-1$ terms are $-1/(2 k^2 + 3 k + 2)$ and $-1/(2 k^2 + 5 k + 3$.
Thus the sum of these two runs is at least



$$ dfrac2k+12 k^2 + 3k + 1 - dfrac2k+22k^2+3k+2 = frac-k(2 k^2 + 3k + 1)(2k^2+3k+2)$$
and at most



$$ dfrac2k+12 k^2 +k + 1 - dfrac2k+22k^2+5k+3 = frac-k(2 k^2 + 3k + 1)(2k^2+3k+2) = frac6k+1(2k+3)(2k^2+k+1)$$



Thus in any case it is bounded in absolute value by $C/k^2$ for some $C$, and thus the series converges.






share|cite|improve this answer













I hope I'm interpreting this correctly: you have one $+$ term, then two $-$, three $+$, four $-$, etc. The start and end of a run of $2k+1$ $+$ terms are $1/(2 k^2 + k + 1)$ and $1/(2 k^2 + 3 k + 1)$ and the start and end of the next run of $2k+2$ $-1$ terms are $-1/(2 k^2 + 3 k + 2)$ and $-1/(2 k^2 + 5 k + 3$.
Thus the sum of these two runs is at least



$$ dfrac2k+12 k^2 + 3k + 1 - dfrac2k+22k^2+3k+2 = frac-k(2 k^2 + 3k + 1)(2k^2+3k+2)$$
and at most



$$ dfrac2k+12 k^2 +k + 1 - dfrac2k+22k^2+5k+3 = frac-k(2 k^2 + 3k + 1)(2k^2+3k+2) = frac6k+1(2k+3)(2k^2+k+1)$$



Thus in any case it is bounded in absolute value by $C/k^2$ for some $C$, and thus the series converges.







share|cite|improve this answer













share|cite|improve this answer



share|cite|improve this answer











answered Sep 15 '17 at 3:16









Robert Israel

304k22201443




304k22201443











  • I get with pari/gp using the shortcut-definition for the binomial-expression x.b = 1+binomial(1+x,2) (so for instance 4.b = 1+binomial(5,2)=11) and using the psi()-function for partial-sums of harmonic numbers for some $N$ for instance N=5;psi(0.b)+(-1)^N*psi(N.b)-2*sum(k=0,N,(-1)^k*psi(k.b)) an alternating sequence of partial sums with decreasing terms. Euler-summation or using sumalt() gives a value of about $ 0.517100379042$
    – Gottfried Helms
    Sep 18 '17 at 10:04
















  • I get with pari/gp using the shortcut-definition for the binomial-expression x.b = 1+binomial(1+x,2) (so for instance 4.b = 1+binomial(5,2)=11) and using the psi()-function for partial-sums of harmonic numbers for some $N$ for instance N=5;psi(0.b)+(-1)^N*psi(N.b)-2*sum(k=0,N,(-1)^k*psi(k.b)) an alternating sequence of partial sums with decreasing terms. Euler-summation or using sumalt() gives a value of about $ 0.517100379042$
    – Gottfried Helms
    Sep 18 '17 at 10:04















I get with pari/gp using the shortcut-definition for the binomial-expression x.b = 1+binomial(1+x,2) (so for instance 4.b = 1+binomial(5,2)=11) and using the psi()-function for partial-sums of harmonic numbers for some $N$ for instance N=5;psi(0.b)+(-1)^N*psi(N.b)-2*sum(k=0,N,(-1)^k*psi(k.b)) an alternating sequence of partial sums with decreasing terms. Euler-summation or using sumalt() gives a value of about $ 0.517100379042$
– Gottfried Helms
Sep 18 '17 at 10:04




I get with pari/gp using the shortcut-definition for the binomial-expression x.b = 1+binomial(1+x,2) (so for instance 4.b = 1+binomial(5,2)=11) and using the psi()-function for partial-sums of harmonic numbers for some $N$ for instance N=5;psi(0.b)+(-1)^N*psi(N.b)-2*sum(k=0,N,(-1)^k*psi(k.b)) an alternating sequence of partial sums with decreasing terms. Euler-summation or using sumalt() gives a value of about $ 0.517100379042$
– Gottfried Helms
Sep 18 '17 at 10:04










up vote
0
down vote













Define $$a_n = displaystylesum_k=fracn(n-1)2 + 1^fracn(n+1)2 frac 1k$$.



So for example, $a_1 = 1$, $a_2 = frac 12 + frac 13$, $a_3= frac 14 + frac 15 + frac 16$, etc.



It's easy to check that series you have written above is $sum_i=1^infty (-1)^i a_i$. It's easy to see that $a_i to 0$,since each fraction in the summation is less than $frac2n(n-1)+ 2$, and there are $n$ of them, so $a_i leq frac2nn(n-1) + 2$, which decreases to zero as $n to infty$.



You can also show that $a_n$ are decreasing as follows :
$a_n - a_n+1 = displaystylesum_k=fracn(n-1)2 + 1^fracn(n+1)2 frac 1k - displaystylesum_k=fracn(n+1)2 + 1^frac(n+2)(n+1)2 frac 1k$



What we do in this sum, is take terms whose difference is exactly $n$. More specifically, I claim that the above sum is the following:
$$
a_n - a_n+1 = frac-2(n+1)(n+2) + displaystylesum_k=fracn(n-1)2 + 1^fracn(n+1)2 left(frac 1k - frac1k+nright)
$$



As if to give an example, $a_3 - a_4 = frac 14 + frac 15 + frac 16 - frac 17 - frac 18 - frac 19 - frac 110 \ = frac 14 - frac 17 + frac 15 - frac 18 + frac 16 - frac 19 - frac 110$. Now you see why this sum holds.



Now, simplify the sums inside :
$$
a_n - a_n+1 = frac-2(n+1)(n+2) + displaystylesum_k=fracn(n-1)2 + 1^fracn(n+1)2 left(fracnk(k+n)right)
$$
Each term on the right hand side is greater than the term with $k = fracn(n+1)2$, and there are $n$ such terms, so:
$$
a_n - a_n+1 geq frac-2(n+1)(n+2) + nleft(fracnfracn(n+1)2left(fracn(n+1)2+nright)right)
$$
which after plenty of simplification becomes:
$$
a_n - a_n+1 geq frac-2(n+1)(n+2) + frac4(n+1)(n+3)
$$
which can be written as:
$$
a_n - a_n+1 geq 4 left(frac1(n+1)(n+3) - frac1(n+1)(2n+2)right)
$$
and the right hand side is positive when $n+3 leq 2n+2 $ i.e. when $n geq 1$.
For example, $a_3 - a_4 $ is predicted to be greater than $frac124$ by this estimate, and it's actual value is $frac3472520$, which is clearly greater than $frac 124$.



So, it follows that $a_n$ is a decreasing sequence which converges to $0$. Now, the alternate series test is applicable, I would think, and this shows that the given series converges conditionally.



Please point out any mistake if found.
$()$






share|cite|improve this answer





















  • Thank you for the acceptance. I would like you to ask somebody to actually compute this sum. I tried doing so on Wolfram Alpha, but it exceeded standard time, so I ran into trouble. That value will settle this question once and for all.
    – Ð°ÑÑ‚он вілла олоф мэллбэрг
    Sep 15 '17 at 3:28














up vote
0
down vote













Define $$a_n = displaystylesum_k=fracn(n-1)2 + 1^fracn(n+1)2 frac 1k$$.



So for example, $a_1 = 1$, $a_2 = frac 12 + frac 13$, $a_3= frac 14 + frac 15 + frac 16$, etc.



It's easy to check that series you have written above is $sum_i=1^infty (-1)^i a_i$. It's easy to see that $a_i to 0$,since each fraction in the summation is less than $frac2n(n-1)+ 2$, and there are $n$ of them, so $a_i leq frac2nn(n-1) + 2$, which decreases to zero as $n to infty$.



You can also show that $a_n$ are decreasing as follows :
$a_n - a_n+1 = displaystylesum_k=fracn(n-1)2 + 1^fracn(n+1)2 frac 1k - displaystylesum_k=fracn(n+1)2 + 1^frac(n+2)(n+1)2 frac 1k$



What we do in this sum, is take terms whose difference is exactly $n$. More specifically, I claim that the above sum is the following:
$$
a_n - a_n+1 = frac-2(n+1)(n+2) + displaystylesum_k=fracn(n-1)2 + 1^fracn(n+1)2 left(frac 1k - frac1k+nright)
$$



As if to give an example, $a_3 - a_4 = frac 14 + frac 15 + frac 16 - frac 17 - frac 18 - frac 19 - frac 110 \ = frac 14 - frac 17 + frac 15 - frac 18 + frac 16 - frac 19 - frac 110$. Now you see why this sum holds.



Now, simplify the sums inside :
$$
a_n - a_n+1 = frac-2(n+1)(n+2) + displaystylesum_k=fracn(n-1)2 + 1^fracn(n+1)2 left(fracnk(k+n)right)
$$
Each term on the right hand side is greater than the term with $k = fracn(n+1)2$, and there are $n$ such terms, so:
$$
a_n - a_n+1 geq frac-2(n+1)(n+2) + nleft(fracnfracn(n+1)2left(fracn(n+1)2+nright)right)
$$
which after plenty of simplification becomes:
$$
a_n - a_n+1 geq frac-2(n+1)(n+2) + frac4(n+1)(n+3)
$$
which can be written as:
$$
a_n - a_n+1 geq 4 left(frac1(n+1)(n+3) - frac1(n+1)(2n+2)right)
$$
and the right hand side is positive when $n+3 leq 2n+2 $ i.e. when $n geq 1$.
For example, $a_3 - a_4 $ is predicted to be greater than $frac124$ by this estimate, and it's actual value is $frac3472520$, which is clearly greater than $frac 124$.



So, it follows that $a_n$ is a decreasing sequence which converges to $0$. Now, the alternate series test is applicable, I would think, and this shows that the given series converges conditionally.



Please point out any mistake if found.
$()$






share|cite|improve this answer





















  • Thank you for the acceptance. I would like you to ask somebody to actually compute this sum. I tried doing so on Wolfram Alpha, but it exceeded standard time, so I ran into trouble. That value will settle this question once and for all.
    – Ð°ÑÑ‚он вілла олоф мэллбэрг
    Sep 15 '17 at 3:28












up vote
0
down vote










up vote
0
down vote









Define $$a_n = displaystylesum_k=fracn(n-1)2 + 1^fracn(n+1)2 frac 1k$$.



So for example, $a_1 = 1$, $a_2 = frac 12 + frac 13$, $a_3= frac 14 + frac 15 + frac 16$, etc.



It's easy to check that series you have written above is $sum_i=1^infty (-1)^i a_i$. It's easy to see that $a_i to 0$,since each fraction in the summation is less than $frac2n(n-1)+ 2$, and there are $n$ of them, so $a_i leq frac2nn(n-1) + 2$, which decreases to zero as $n to infty$.



You can also show that $a_n$ are decreasing as follows :
$a_n - a_n+1 = displaystylesum_k=fracn(n-1)2 + 1^fracn(n+1)2 frac 1k - displaystylesum_k=fracn(n+1)2 + 1^frac(n+2)(n+1)2 frac 1k$



What we do in this sum, is take terms whose difference is exactly $n$. More specifically, I claim that the above sum is the following:
$$
a_n - a_n+1 = frac-2(n+1)(n+2) + displaystylesum_k=fracn(n-1)2 + 1^fracn(n+1)2 left(frac 1k - frac1k+nright)
$$



As if to give an example, $a_3 - a_4 = frac 14 + frac 15 + frac 16 - frac 17 - frac 18 - frac 19 - frac 110 \ = frac 14 - frac 17 + frac 15 - frac 18 + frac 16 - frac 19 - frac 110$. Now you see why this sum holds.



Now, simplify the sums inside :
$$
a_n - a_n+1 = frac-2(n+1)(n+2) + displaystylesum_k=fracn(n-1)2 + 1^fracn(n+1)2 left(fracnk(k+n)right)
$$
Each term on the right hand side is greater than the term with $k = fracn(n+1)2$, and there are $n$ such terms, so:
$$
a_n - a_n+1 geq frac-2(n+1)(n+2) + nleft(fracnfracn(n+1)2left(fracn(n+1)2+nright)right)
$$
which after plenty of simplification becomes:
$$
a_n - a_n+1 geq frac-2(n+1)(n+2) + frac4(n+1)(n+3)
$$
which can be written as:
$$
a_n - a_n+1 geq 4 left(frac1(n+1)(n+3) - frac1(n+1)(2n+2)right)
$$
and the right hand side is positive when $n+3 leq 2n+2 $ i.e. when $n geq 1$.
For example, $a_3 - a_4 $ is predicted to be greater than $frac124$ by this estimate, and it's actual value is $frac3472520$, which is clearly greater than $frac 124$.



So, it follows that $a_n$ is a decreasing sequence which converges to $0$. Now, the alternate series test is applicable, I would think, and this shows that the given series converges conditionally.



Please point out any mistake if found.
$()$






share|cite|improve this answer













Define $$a_n = displaystylesum_k=fracn(n-1)2 + 1^fracn(n+1)2 frac 1k$$.



So for example, $a_1 = 1$, $a_2 = frac 12 + frac 13$, $a_3= frac 14 + frac 15 + frac 16$, etc.



It's easy to check that series you have written above is $sum_i=1^infty (-1)^i a_i$. It's easy to see that $a_i to 0$,since each fraction in the summation is less than $frac2n(n-1)+ 2$, and there are $n$ of them, so $a_i leq frac2nn(n-1) + 2$, which decreases to zero as $n to infty$.



You can also show that $a_n$ are decreasing as follows :
$a_n - a_n+1 = displaystylesum_k=fracn(n-1)2 + 1^fracn(n+1)2 frac 1k - displaystylesum_k=fracn(n+1)2 + 1^frac(n+2)(n+1)2 frac 1k$



What we do in this sum, is take terms whose difference is exactly $n$. More specifically, I claim that the above sum is the following:
$$
a_n - a_n+1 = frac-2(n+1)(n+2) + displaystylesum_k=fracn(n-1)2 + 1^fracn(n+1)2 left(frac 1k - frac1k+nright)
$$



As if to give an example, $a_3 - a_4 = frac 14 + frac 15 + frac 16 - frac 17 - frac 18 - frac 19 - frac 110 \ = frac 14 - frac 17 + frac 15 - frac 18 + frac 16 - frac 19 - frac 110$. Now you see why this sum holds.



Now, simplify the sums inside :
$$
a_n - a_n+1 = frac-2(n+1)(n+2) + displaystylesum_k=fracn(n-1)2 + 1^fracn(n+1)2 left(fracnk(k+n)right)
$$
Each term on the right hand side is greater than the term with $k = fracn(n+1)2$, and there are $n$ such terms, so:
$$
a_n - a_n+1 geq frac-2(n+1)(n+2) + nleft(fracnfracn(n+1)2left(fracn(n+1)2+nright)right)
$$
which after plenty of simplification becomes:
$$
a_n - a_n+1 geq frac-2(n+1)(n+2) + frac4(n+1)(n+3)
$$
which can be written as:
$$
a_n - a_n+1 geq 4 left(frac1(n+1)(n+3) - frac1(n+1)(2n+2)right)
$$
and the right hand side is positive when $n+3 leq 2n+2 $ i.e. when $n geq 1$.
For example, $a_3 - a_4 $ is predicted to be greater than $frac124$ by this estimate, and it's actual value is $frac3472520$, which is clearly greater than $frac 124$.



So, it follows that $a_n$ is a decreasing sequence which converges to $0$. Now, the alternate series test is applicable, I would think, and this shows that the given series converges conditionally.



Please point out any mistake if found.
$()$







share|cite|improve this answer













share|cite|improve this answer



share|cite|improve this answer











answered Sep 15 '17 at 3:18









астон вілла олоф мэллбэрг

32k22463




32k22463











  • Thank you for the acceptance. I would like you to ask somebody to actually compute this sum. I tried doing so on Wolfram Alpha, but it exceeded standard time, so I ran into trouble. That value will settle this question once and for all.
    – Ð°ÑÑ‚он вілла олоф мэллбэрг
    Sep 15 '17 at 3:28
















  • Thank you for the acceptance. I would like you to ask somebody to actually compute this sum. I tried doing so on Wolfram Alpha, but it exceeded standard time, so I ran into trouble. That value will settle this question once and for all.
    – Ð°ÑÑ‚он вілла олоф мэллбэрг
    Sep 15 '17 at 3:28















Thank you for the acceptance. I would like you to ask somebody to actually compute this sum. I tried doing so on Wolfram Alpha, but it exceeded standard time, so I ran into trouble. That value will settle this question once and for all.
– Ð°ÑÑ‚он вілла олоф мэллбэрг
Sep 15 '17 at 3:28




Thank you for the acceptance. I would like you to ask somebody to actually compute this sum. I tried doing so on Wolfram Alpha, but it exceeded standard time, so I ran into trouble. That value will settle this question once and for all.
– Ð°ÑÑ‚он вілла олоф мэллбэрг
Sep 15 '17 at 3:28


這個網誌中的熱門文章

How to combine Bézier curves to a surface?

Mutual Information Always Non-negative

Why am i infinitely getting the same tweet with the Twitter Search API?