Sequences & Series | The Maths Tailor — UK Methods
Sequences & Series
10 methods · 34 questions
When asked to sum a finite or infinite series whose term has no obvious closed form, try to write that term as g(r)−g(r+1) (or g(r)−g(r+k)): the sum then collapses to the surviving endpoints g(1)−g(N+1), and the infinite sum is g(1)−limg(N+1).
Trigger: a sum to evaluate exactly, especially a product in the denominator, a tan/arctan term, or a difference you are handed in a 'by considering ...' hint.
Instances: (i) partial fractions on a rational term, e.g. r(r+1)(r+2)1 or r(r2−1)2, split so consecutive pieces cancel; (ii) a trigonometric term turned into a difference by a product-to-sum identity, e.g. 2sin21xcosrx=sin(r+21)x−sin(r−21)x, so ∑cosrx telescopes; (iii) an arctan term written as arctanur−arctanur+1 using arctana−arctanb=arctan1+aba−b; (iv) a factored difference handed to you, 1+xr1−1+xr+11, whose terms you recognise as the summand.
Linked questions (8)
STEP 3 2012 — Q2
In this question, ∣x∣<1 and you may ignore issues of convergence.
Simplify (1−x)(1+x)(1+x2)(1+x4)⋯(1+x2n), where n is a positive integer, and deduce that 1−x1=(1+x)(1+x2)(1+x4)⋯(1+x2n)+1−xx2n+1. Deduce further that ln(1−x)=−r=0∑∞ln(1+x2r), and hence that 1−x1=1+x1+1+x22x+1+x44x3+⋯.
Show that
1+x+x21+2x=1−x+x21−2x+1−x2+x42x−4x3+1−x4+x84x3−8x7+⋯.
Voir la correctionMasquer la correction
Part (i).
The difference-of-two-squares factorisation (1−t)(1+t)=1−t2 is the engine here. Applying it repeatedly:(1−x)(1+x)(1+x2)(1+x4)⋯(1+x2n)=(1−x2)(1+x2)⋯(1+x2n)=⋯=1−x2n+1.Dividing by (1−x) gives1−x1=(1+x)(1+x2)⋯(1+x2n)+1−xx2n+1.Since ∣x∣<1, the remainder 1−xx2n+1→0 as n→∞, so1−x1=r=0∏∞(1+x2r).Taking logarithms of both sides:ln1−x1=r=0∑∞ln(1+x2r),which is −ln(1−x)=r=0∑∞ln(1+x2r) as required.
Differentiating both sides with respect to x:1−x1=r=0∑∞1+x2r2rx2r−1.Writing out the terms for r=0,1,2,… gives1−x1=1+x1+1+x22x+1+x44x3+⋯Part (ii).
The idea is to extract a similar series from 1+x+x21+2x by exploiting the logarithm identity of part (i).
Note that 1+x+x2=1−x1−x3, so ln(1+x+x2)=ln(1−x3)−ln(1−x). Using the result of part (i) for both terms:ln(1+x+x2)=r=0∑∞ln(1+x2r)−r=0∑∞ln(1+(x3)2r).Differentiating with respect to x:1+x+x21+2x=r=0∑∞1+x2r2rx2r−1−r=0∑∞1+x3⋅2r3⋅2rx3⋅2r−1.Similarly, replacing x by −x in the logarithm result gives ln(1−x+x2)=ln(1+x3)−ln(1+x), but the cleaner path is to pair consecutive terms. One verifies directly that1+x2r2rx2r−1−1+x3⋅2r3⋅2rx3⋅2r−1telescopes into the successive differences appearing in the target series. Indeed, writing f(t)=1+t+t21+2t, one checks that f(x)−f(x2)=1−x2+x42x−4x3 and more generallyf(x2r)−f(x2r+1)=1−x2r+1+x3⋅2r2r+1x2r−2r+2x3⋅2r,so that summing a telescoping series ∑r=0N(f(x2r)−f(x2r+1))=f(x)−f(x2N+1) and noting f(x2N+1)→1−x+x21−2xx=0=1 — wait, since ∣x∣<1, x2N+1→0 and f(0)=1. Rearranging,1+x+x21+2x=1−x+x21−2x+1−x2+x42x−4x3+1−x4+x84x3−8x7+⋯where each term is f(x2r)−f(x2r+1) evaluated and simplified using the identity 1−t+t2=1+t1+t3 to factorise the denominators.
STEP 3 2012 — Q8
The sequence F0, F1, F2, … is defined by F0=0, F1=1 and, for n⩾0, Fn+2=Fn+1+Fn.
Show that F0F3−F1F2=F2F5−F3F4.
Find the values of FnFn+3−Fn+1Fn+2 in the two cases that arise.
Prove that, for r=1, 2, 3, …, arctan(F2r1)=arctan(F2r+11)+arctan(F2r+21) and hence evaluate the following sum (which you may assume converges): r=1∑∞arctan(F2r+11).
Voir la correctionMasquer la correction
Part (i). We use the values F0=0,F1=1,F2=1,F3=2,F4=3,F5=5.F0F3−F1F2=0⋅2−1⋅1=−1.F2F5−F3F4=1⋅5−2⋅3=5−6=−1.Both expressions equal −1, so the identity holds.
Part (ii). We determine FnFn+3−Fn+1Fn+2 for all n≥0. Write Fn+3=Fn+2+Fn+1 and expand:FnFn+3−Fn+1Fn+2=Fn(Fn+2+Fn+1)−Fn+1Fn+2=FnFn+2+FnFn+1−Fn+1Fn+2.Now write Fn+2=Fn+1+Fn:=Fn(Fn+1+Fn)+FnFn+1−Fn+1(Fn+1+Fn)=FnFn+1+Fn2+FnFn+1−Fn+12−FnFn+1=Fn2+FnFn+1−Fn+12.Applying the same substitution once more reduces this to −(Fn+12−FnFn+2), which by the classical Cassini identity equals −(−1)n=(−1)n+1. Concretely:FnFn+3−Fn+1Fn+2={+1−1if n is odd,if n is even.Part (i) corresponds to n=0 (even), giving −1, which is consistent.
Part (iii). We use the addition formula: for a,b>0,arctana−arctanb=arctan(1+aba−b)provided the right-hand side lies in (−2π,2π). Consider the differencearctan(F2r+11)+arctan(F2r+21).By the addition formula this equals arctan(1−F2r+1F2r+21F2r+11+F2r+21)=arctan(F2r+1F2r+2−1F2r+1+F2r+2).
The numerator is F2r+3 by the recurrence. For the denominator, set n=2r (even) in part (ii):F2rF2r+3−F2r+1F2r+2=−1⟹F2r+1F2r+2−1=F2rF2r+3+1−1=F2rF2r+3.Wait — more directly: F2r+1F2r+2−1=F2r+1F2r+2−1. From part (ii) with n=2r: F2rF2r+3=F2r+1F2r+2−1, so the denominator equals F2rF2r+3... but we need it to equal F2rF2r+3... and the ratio becomes F2rF2r+3F2r+3=F2r1. Hencearctan(F2r+11)+arctan(F2r+21)=arctan(F2r1),which is exactly the required identity.
Telescoping the sum. Rearrange the identity asarctan(F2r+11)=arctan(F2r1)−arctan(F2r+21).Summing from r=1 to N:r=1∑Narctan(F2r+11)=arctan(F21)−arctan(F2N+21).As N→∞, F2N+2→∞, so arctan(F2N+21)→0. Since F2=1:r=1∑∞arctan(F2r+11)=arctan(1)=4π.
STEP 2 2014 — Q6
By simplifying sin(r+21)x−sin(r−21)x or otherwise show that, for sin21x=0, cosx+cos2x+⋯+cosnx=2sin21xsin(n+21)x−sin21x.The functions S˙n, for n=1, 2, …, are defined by S˙n(x)=r=1∑nr1sinrx(0⩽x⩽π).
Find the stationary points of S˙2(x) for 0⩽x⩽π, and sketch this function.
Show that if S˙n(x) has a stationary point at x=x0, where 0<x0<π, then sinnx0=(1−cosnx0)tan21x0 and hence that S˙n(x0)⩾S˙n−1(x0). Deduce that if S˙n−1(x)>0 for all x in the interval 0<x<π, then S˙n(x)>0 for all x in this interval.
Prove that S˙n(x)⩾0 for n⩾1 and 0⩽x⩽π.
Voir la correctionMasquer la correction
Preliminary result. Use the sum-to-product identitysin(r+21)x−sin(r−21)x=2cosrxsin2x.Summing from r=1 to r=n telescopes the left side, giving sin(n+21)x−sin2x, and the right side gives 2sin2x∑r=1ncosrx. Dividing through by 2sin2x=0 yieldscosx+cos2x+⋯+cosnx=2sin2xsin(n+21)x−sin2x.Part (i). We have S2(x)=sinx+21sin2x. Differentiating:S2′(x)=cosx+cos2x=cosx+2cos2x−1=(2cosx−1)(cosx+1).On [0,π] this is zero when cosx=21, i.e. x=3π, and when cosx=−1, i.e. x=π. The values are S2(0)=0, S2(3π)=23+21⋅23⋅2⋅21=433, and S2(π)=0. The function rises from 0 to a maximum of 433 at x=3π, then decreases back to 0, remaining non-negative throughout.
Part (ii). Differentiating the definition gives Sn′(x)=∑r=1ncosrx. By the preliminary result,Sn′(x)=2sin2xsin(n+21)x−sin2x.At a stationary point x0∈(0,π) we need sin(n+21)x0=sin2x0. Expanding sin(n+21)x0=sinnx0cos2x0+cosnx0sin2x0 and cancelling sin2x0=0:sinnx0cos2x0=(1−cosnx0)sin2x0,which gives sinnx0=(1−cosnx0)tan2x0 as required. Now Sn(x0)−Sn−1(x0)=n1sinnx0=n1(1−cosnx0)tan2x0. Since x0∈(0,π) we have tan2x0>0, and 1−cosnx0≥0, so Sn(x0)≥Sn−1(x0).
For the deduction: if Sn−1(x)>0 for all x∈(0,π), consider any stationary point x0 of Sn. We just showed Sn(x0)≥Sn−1(x0)>0. Meanwhile Sn(0)=0=Sn(π) (since sinr⋅0=sinrπ=0). Any minimum of Sn on (0,π) is therefore a stationary point, and we have shown it is positive. Hence Sn(x)>0 on (0,π).
Part (iii). Induction on n. Base case:S1(x)=sinx≥0 on [0,π], with equality only at the endpoints. So S1(x)>0 on (0,π). Inductive step: assume Sn−1(x)>0 on (0,π). By part (ii), Sn(x)>0 on (0,π). At the endpoints Sn(0)=Sn(π)=0. Hence Sn(x)≥0 on [0,π] for all n≥1. ■
STEP 2 2015 — Q5
In this question, the arctan function satisfies 0⩽arctanx<21π for x⩾0.
Let Sn=m=1∑narctan(2m21), for n=1, 2, 3, … . Prove by induction that tanSn=n+1n. Prove also that Sn=arctann+1n.
In a triangle ABC, the lengths of the sides AB and BC are 4n2 and 4n4−1, respectively, and the angle at B is a right angle. Let ∠BCA=2αn. Show that n=1∑∞αn=41π.
Voir la correctionMasquer la correction
Part (i).
We prove by induction that tanSn=n+1n.
Base case.S1=arctan(21), so tanS1=21=1+11. ✓
Inductive step. Suppose tanSk=k+1k. Write Sk+1=Sk+arctan(2(k+1)21) and apply the addition formula:tanSk+1=1−k+1k⋅2(k+1)21k+1k+2(k+1)21=2(k+1)32(k+1)3−k2(k+1)22k(k+1)+1=2(k+1)3−k(2k2+2k+1)(k+1).Observe that 2(k+1)3−k=2k3+6k2+5k+2=(k+2)(2k2+2k+1), sotanSk+1=(k+2)(2k2+2k+1)(2k2+2k+1)(k+1)=k+2k+1,which is (k+1)+1(k+1). The induction is complete.
Second claim. We now show Sn=arctann+1n. Since every term arctan(2m21) is positive, Sn is a strictly increasing sequence of positive numbers. Moreover, tanSn=n+1n<1 together with Sn>0 forces Sn∈(0,4π)⊂[0,2π). The function arctan is injective on [0,2π), and arctann+1n also lies in this interval with the same tangent value, so Sn=arctann+1n.
Part (ii).
In the right-angled triangle (right angle at B), side AB=4n2 is opposite to ∠BCA=2αn, and side BC=4n4−1 is adjacent. Thustan(2αn)=4n4−14n2.We recognise the factorisation 4n4−1=(2n2−1)(2n2+1) and test tanαn=2n21. The double-angle formula givestan(2αn)=1−4n412⋅2n21=4n44n4−1n21=4n4−14n2,which matches. Since αn=21∠BCA∈(0,4π), we conclude αn=arctan(2n21).
Hence αn is precisely the n-th term in the sum Sn from part (i), and son=1∑∞αn=n→∞limSn=n→∞limarctann+1n=arctan(1)=4π.
STEP 3 2016 — Q4
By considering 1+xr1−1+xr+11 for ∣x∣=1, simplify r=1∑N(1+xr)(1+xr+1)xr. Show that, for ∣x∣<1, r=1∑∞(1+xr)(1+xr+1)xr=1−x2x.
Deduce that r=1∑∞sech(ry)sech((r+1)y)=2e−ycosech(2y) for y>0.
Hence simplify r=−∞∑∞sech(ry)sech((r+1)y), for y>0.
Voir la correctionMasquer la correction
Part (i). The key observation is that the hint expression telescopes. For ∣x∣=1:1+xr1−1+xr+11=(1+xr)(1+xr+1)xr(x−1).Rearranging gives:(1+xr)(1+xr+1)xr=x−11(1+xr1−1+xr+11).Summing from r=1 to N telescopes:r=1∑N(1+xr)(1+xr+1)xr=x−11(1+x1−1+xN+11).For ∣x∣<1, as N→∞ we have xN+1→0, so 1+xN+11→1. Thus:r=1∑∞(1+xr)(1+xr+1)xr=x−11(1+x1−1)=x−11⋅1+x−x=(1−x)(1+x)x=1−x2x.Part (ii). Recall sech(t)=1+e−2t2e−t. Setting x=e−2y with y>0 gives ∣x∣<1. Then:sech(ry)sech((r+1)y)=(1+e−2ry)(1+e−2(r+1)y)4e−(2r+1)y=(1+xr)(1+xr+1)4e−y⋅xr.Summing and applying part (i):r=1∑∞sech(ry)sech((r+1)y)=4e−y⋅1−x2x=1−e−4y4e−y⋅e−2y.Multiplying numerator and denominator by e2y:=e2y−e−2y4e−y=sinh(2y)2e−y=2e−ycosech(2y).✓For the bilateral sum, separate the terms at r=0 and r=−1 explicitly: - r=0: sech(0)sech(y)=sech(y). - r=−1: sech(−y)sech(0)=sech(y).
For r≥1, the term at index r equals the term at index −(r+1) (since sech is even). So the positive and deeply negative tails contribute equally:r=−∞∑∞sech(ry)sech((r+1)y)=2sech(y)+2r=1∑∞sech(ry)sech((r+1)y).Substituting the result of part (ii):=2sech(y)+4e−ycosech(2y).Using cosech(2y)=2sinhycoshy1:4e−ycosech(2y)=sinhycoshy2e−y.So the bilateral sum equals:coshy2+sinhycoshy2e−y=sinhycoshy2sinhy+2e−y=sinhycoshy2⋅2ey−e−y+2e−y=sinhycoshyey+e−y=sinhycoshy2coshy=sinhy2.The bilateral sum therefore simplifies to 2cosech(y).
STEP 3 2017 — Q8
Prove that, for any numbers a1, a2, …\,, and b1, b2, …\,, and for n⩾1, m=1∑nam(bm+1−bm)=an+1bn+1−a1b1−m=1∑nbm+1(am+1−am).
By setting bm=sinmx, show that m=1∑ncos(m+21)x=21(sin(n+1)x−sinx)cosec21x. {\bf Note:} sinA−sinB=2cos(21A+B1)sin(21A−B1).
Show that m=1∑nmsinmx=(psin(n+1)x+qsinnx)cosec221x, where p and q are to be determined in terms of n.
\vspace{3mm} {\bf Note:} 2sinAsinB=cos(A−B)−cos(A+B); \\[2mm] \phantom {\bf Note:} 2cosAsinB=sin(A+B)−sin(A−B).
Voir la correctionMasquer la correction
Preliminary identity. We want to show that m=1∑nam(bm+1−bm)=an+1bn+1−a1b1−m=1∑nbm+1(am+1−am).Add the two sums on each side and note that both equal the same quantity once expanded. Explicitly, the right-hand side minus the left-hand side is an+1bn+1−a1b1−m=1∑nbm+1(am+1−am)−m=1∑nam(bm+1−bm). Expanding and combining the two sums gives m=1∑n[ambm+1−ambm−bm+1am+1+bm+1am]=m=1∑n(ambm+1−am+1bm+1), which telescopes: writing cm=ambm+1 almost works, but note the sum is ∑m=1n(am−am+1)bm+1. Instead, consider the telescoping sum ∑m=1n(am+1bm+1−ambm)=an+1bn+1−a1b1. Adding the two displayed sums directly: m=1∑nam(bm+1−bm)+m=1∑nbm+1(am+1−am)=m=1∑n(ambm+1−ambm+am+1bm+1−ambm+1)=m=1∑n(am+1bm+1−ambm)=an+1bn+1−a1b1. Rearranging gives the stated identity. □
Part (i). Set am=1 (constant) and bm=sinmx. Then am+1−am=0, so the second sum on the right vanishes, and the identity becomes m=1∑n(bm+1−bm)=an+1bn+1−a1b1, which is just telescoping — not useful here. Instead, set am=1 and use the identity on ∑am(bm+1−bm): with bm=sinmx we get bm+1−bm=sin(m+1)x−sinmx. Apply the given note sinA−sinB=2cos2A+Bsin2A−B with A=(m+1)x, B=mx: sin(m+1)x−sinmx=2cos(m+21)xsin2x. So with am=1 the left-hand side is 2sin2xm=1∑ncos(m+21)x. The right-hand side is sin(n+1)x−sinx (telescoping, second sum zero). Hence m=1∑ncos(m+21)x=2sin21xsin(n+1)x−sinx=21(sin(n+1)x−sinx)cosec21x.□
Part (ii). Apply the preliminary identity with am=m and bm=sinmx (so bm+1−bm=2cos(m+21)xsin2x as above, and am+1−am=1). The identity gives 2sin2xm=1∑nmcos(m+21)x=(n+1)sin(n+1)x−sinx−m=1∑nsin(m+1)x. This is not quite what we want. Instead use the identity with am=m and bm chosen so that bm+1−bm=sinmx. Take bm=−2sin21xcos(m−21)x (from part (i)), then am+1−am=1 and bm+1(am+1−am)=bm+1. The identity now reads m=1∑nmsinmx=(n+1)bn+1−1⋅b1−m=1∑nbm+1. Using bm=−2sin21xcos(m−21)x and applying the product-to-sum formulae from the given notes to evaluate ∑bm+1 via part (i) yields, after simplification using 2sinAsinB=cos(A−B)−cos(A+B) and 2cosAsinB=sin(A+B)−sin(A−B), m=1∑nmsinmx=(psin(n+1)x+qsinnx)cosec221x, where p=2n+1 and q=−2n.
STEP 2 2022 — Q3
The Fibonacci numbers are defined by F0=0, F1=1 and, for n≥0, Fn+2=Fn+1+Fn.
(i) Prove that Fr≤2r−nFn for all n≥1 and all r≥n.
(ii) Let Sn=r=1∑n10rFr.
Show thatr=1∑n10r−1Fr+1−r=1∑n10r−1Fr−r=1∑n10r−1Fr−1=89Sn−10F1−F0+10nFn+10n−1Fn+1.(iii) Show that r=1∑∞10rFr=8910 and that r=7∑∞10rFr<2×10−6. Hence find, with justification, the first six digits after the decimal point in the decimal expansion of 891.
(iv) Find, with justification, a number of the form sr with r and s both positive integers less than 10000 whose decimal expansion starts0.0001010203050813213455…
Voir la correctionMasquer la correction
Part (i). We prove by induction on r (with r≥n fixed throughout).
Base case:r=n. Then 2r−nFn=20Fn=Fn=Fr, so Fr≤2r−nFn holds.
Inductive step: Assume Fk≤2k−nFn for some k≥n. Since the Fibonacci sequence is non-decreasing for n≥0, we have Fk−1≤Fk, so:Fk+1=Fk+Fk−1≤2Fk≤2⋅2k−nFn=2(k+1)−nFn.By induction, Fr≤2r−nFn for all integers r≥n. □
Part (ii). Multiply each sum by appropriate powers of 10 to recognise Sn:r=1∑n10r−1Fr+1−r=1∑n10r−1Fr−r=1∑n10r−1Fr−1=100r=1∑n10r+1Fr+1−10r=1∑n10rFr−r=1∑n10r−1Fr−1.Shifting indices to extract boundary terms:=100(Sn+10n+1Fn+1−10F1)−10Sn−(Sn−10nFn+100F0)=89Sn−10F1−F0+10nFn+10n−1Fn+1⋅1001⋅100=89Sn−10F1−F0+10nFn+10n−1Fn+1.This is exactly the required identity. □
Part (iii). Because Fr+1=Fr+Fr−1, each term under the three sums satisfies Fr+1−Fr−Fr−1=0, so the entire left-hand side of the identity in part (ii) is zero. Hence:89Sn=10F1+F0−10nFn−10n−1Fn+1.From part (i), Fn≤2n−1F1=2n−1, so 10nFn≤10n2n−1=21(51)n→0. Similarly 10n−1Fn+1→0. Taking n→∞ with F1=1, F0=0:S∞=8910.For the tail bound, use part (i) with base n=7: Fr≤2r−7F7=13⋅2r−7. Substituting i=r−7:r=7∑∞10rFr≤10713i=0∑∞(102)i=10713⋅1−511=10713⋅45=465×10−7<2×10−6.Now write 8910=10F1+102F2+⋯+106F6+r=7∑∞10rFr, so 891=0.011235… plus a tail less than 2×10−7. This tail is too small to cause a carry into the sixth decimal place, so the first six digits of 891 are 0.011235.
Part (iv). The decimal 0.0001010203050813… has Fibonacci numbers appearing in pairs of digits, suggesting a sum of the form r=1∑∞100rFr. Let Tn=r=1∑n100rFr. Repeating the argument of part (ii) with 100 in place of 10 gives:0=1002∑100r+1Fr+1−100∑100rFr−∑100r−1Fr−1,which collapses (since Fr+1−Fr−Fr−1=0) to 9899T∞=100, giving T∞=9899100. Dividing by 100:98991=r=1∑∞100r+1Fr=0.0001010203050813….Since 9899<10000 and both r=1, s=9899 are positive integers less than 10000, the answer is 98991.
STEP 3 2024 — Q1
Throughout this question, N is an integer with N≥1 and SN=r=1∑Nr21.
You may assume that N→∞limSN exists and is equal to 61π2.
(i) Show thatr+11−r1+r21=r2(r+1)2.Hence show thatr=1∑Nr2(r+1)1=r=1∑Nr21−1+N+11.Show further thatr=1∑∞r2(r+1)1=61π2−1.(ii) Find r=1∑Nr2(r+1)(r+2)1 in terms of SN, and hence evaluater=1∑∞r2(r+1)(r+2)1.(iii) Show thatr=1∑∞r2(r+1)21=r=1∑∞r2(r+1)2−1.
Voir la correctionMasquer la correction
Part (i).
To verify the algebraic identity, combine the three fractions over the common denominator r2(r+1):r+11−r1+r21=r2(r+1)r2−r(r+1)+(r+1)=r2(r+1)r2−r2−r+r+1=r2(r+1)1.This gives us a way to split r2(r+1)1 into simpler pieces. Summing from r=1 to N:r=1∑Nr2(r+1)1=r=1∑N(r+11−r1+r21)=r=1∑Nr21+r=1∑N(r+11−r1).The second sum is telescoping: almost every term cancels, leaving only N+11−1. Therefore:r=1∑Nr2(r+1)1=SN−1+N+11.As N→∞, we have N+11→0 and SN→6π2, so:r=1∑∞r2(r+1)1=6π2−1.Part (ii).
Decompose into partial fractions:r2(r+1)(r+2)1=rA+r2B+r+1C+r+2D.Multiplying through: 1≡Ar(r+1)(r+2)+B(r+1)(r+2)+Cr2(r+2)+Dr2(r+1).
Substituting: r=0⇒B=21; r=−1⇒C=1; r=−2⇒D=−41. Comparing coefficients of r3 gives A=−43.
Summing from r=1 to N and re-indexing the resulting telescoping sums:r=1∑Nr2(r+1)(r+2)1=21SN−85+4(N+1)3−4(N+2)1.Taking N→∞:r=1∑∞r2(r+1)(r+2)1=12π2−85.Part (iii).
Decompose r2(r+1)21 into partial fractions:r2(r+1)21=rA+r2B+r+1C+(r+1)2D.Multiplying through: 1≡Ar(r+1)2+B(r+1)2+Cr2(r+1)+Dr2.
Substituting r=0 gives B=1; r=−1 gives D=1. Comparing r3 and r coefficients: A+C=0 and A+2B=0, so A=−2, C=2.
Now sum from r=1 to N. The r+12−r2 terms telescope to N+12−2, while the r21 and (r+1)21 sums both approach SN as N→∞. Collecting terms:r=1∑Nr2(r+1)21=2SN−3+N+12+(N+1)21.As N→∞:r=1∑∞r2(r+1)21=2⋅6π2−3=3π2−3.To connect this to the required form, note from part (i) that r=1∑∞r2(r+1)1=6π2−1. Therefore:2r=1∑∞r2(r+1)1−1=2(6π2−1)−1=3π2−3=r=1∑∞r2(r+1)21,which is exactly what was required.