ADVANCED PROBLEMS AND SOLUTIONS

EDITED BY
FLORIAN LUCA

Please send all communications concerning ADVANCED PROBLEMS AND SOLUTIONS

to FLORIAN LUCA, SCHOOL OF MATHEMATICS, UNIVERSITY OF THE WITWA-
TERSRAND, PRIVATE BAG X3, WITS 2050, JOHANNESBURG, SOUTH AFRICA or by
e-mail at the address [email protected] as files of the type tex, dvi, ps, doc, html, pdf,
etc. This department especially welcomes problems believed to be new or extending old results.
Proposers should submit solutions or other information that will assist the editor. To facilitate
their consideration, all solutions sent by regular mail should be submitted on separate signed
sheets within two months after publication of the problems.

PROBLEMS PROPOSED IN THIS ISSUE

H-809 Proposed by Hideyuki Ohtsuka, Saitama, Japan.

Prove that
     √
α β α β 7 5−5
1− 1− 1− 1− ··· = .
L2 L22 L23 L24 22

H-810 Proposed by Ángel Plaza, Gran Canaria, Spain.

Prove that
∞
X 1 5 1
4
= − √ .
L − 25
n=3 n
63 6 5

H-811 Proposed by Ángel Plaza, Gran Canaria, Spain.

For any positive integer k let {Fk,n }n≥0 be defined by Fk,n+2 = kFk,n+1 + Fk,n for n ≥ 0
with Fk,0 = 0, Fk,1 = 1. Prove that
∞ √
X 1 k2 + 4
= .
n=0
1 + Fk,2n+1 2k

H-812 Proposed by Hideyuki Ohtsuka, Saitama, Japan.

Prove that
X Fn+1 Fn Ln  X Fn Ln  2j 
= .
i i j i i Fn+1
i+j=Fn+1 i+j=Fn+1

282 VOLUME 55, NUMBER 3

 ADVANCED PROBLEMS AND SOLUTIONS

SOLUTIONS

A Series Related to the Sum of the Reciprocals of the Fibonacci Numbers

H-775 Proposed by H. Ohtsuka, Saitama, Japan.

(Vol. 53, No. 3, August 2015)
Let c be any real number c 6= 2, −L2n for n ≥ 0. Let
∞ 
√ Y c −1

γc = 5 1+ .
L2n
n=1
Prove that
∞
X 1 γc + c − 3
= 2 .
(L2 + c)(L4 + c) · · · (L2k + c) c −c−2
k=1

Solution by the proposer.

Let Pn = (L2 + c)(L22 + c) · · · (L2n + c). For n ≥ 1, we show that
n
X 1 L n+1 − c
(c2 − c − 2) = 2 + c − 3. (1)
Pk Pn
k=1
The proof of (1) is by mathematical induction on n. For n = 1, we have
c2 − c − 2 c2 − c − 2 7−c L4 − c
LHS = = = +c−3 = + c − 3 = RHS.
P1 3+c 3+c P1
We assume that (1) holds for n. For n + 1, we have
n+1 n
!
X 1 1 X 1
(c2 − c − 2) = (c2 − c − 2) +
Pk Pn+1 Pk
k=1 k=1
c2
− c − 2 L2n+1 − c
= + +c−3
Pn+1 Pn
c2 − c − 2 + (L2n+1 − c)(L2n+1 + c)
= +c−3
Pn+1
L2n+1 − 2 − c
= 2 +c−3
Pn+1
L n+2 − c
= 2 + c − 3,
Pn+1
since L2m − 2(−1)m = L2m . Thus, (1) holds for n + 1. Therefore (1) is proved. We have
n n n  
Y Y Y c
Pn = (L2k + c) = L2k 1+ .
L2k
k=1 k=1 k=1
Hence, using Fm Lm = F2m , we have
F2 L2 L4 L8 · · · L2n = F4 L4 L8 · · · L2n = · · · = F2n L2n = F2n+1 .
Thus,
n  
Y c
Pn = F2n+1 1+ . (2)
L2k
k=1

AUGUST 2017 283

THE FIBONACCI QUARTERLY

Therefore, by (1) and (2), we have

n n 
" #
c −1

X 1 1 L2n+1 − c Y γc + c − 3
= 2 1+ +c−3 → 2
Pk c −c−2 F2n+1 L2k c −c−2
k=1 k=1
√
as n → ∞ since Lm /Fm → 5 as m → ∞.
Note: If c = 0, we then have
∞ ∞ √
X 1 γ0 − 3 X 1 3− 5
= i.e., = .
L2 L4 · · · L2k −2 F2k+1 2
k=1 k=1

From the above identity, we obtain the well-known identity

∞ √
X 1 7− 5
= .
F2k 2
k=0

Also solved by Dmitry Fleischman.

A Series of Inverse Tangents of Reciprocals of Lucas Numbers

H-776 Proposed by H. Ohtsuka, Saitama, Japan.

(Vol. 53, No. 3, August 2015)
Determine
∞ ∞
X 1 X 1 1
(i) (−1)n tan−1 and (ii) tan−1 tan−1 .
n=0
L3n n=1
F2n L2n

Solution by the proposer.

(i) For n ≥ 0, we have
1 1
!
1 1 α3n
+
−1 −1 α3n+1
tan n + tan = tan−1 1 1
α3 α n+1
3 1− α3n
· 3n+1
α
n+1 n
!
−1 α3 + α3
= tan n
α4·3 − 1
!
2·3n α3n + α−3n


α
= tan−1
α2·3n (α3n + α−3n ) (α3n − α−3n )
 
−1 1 1
= tan 3n 3n = tan−1 .
α +β L3n
Using the above identity, we have
m m  
X
n −1 1 X
n −1 1 n+1 −1 1
(−1) tan = (−1) tan − (−1) tan
L3n α3n α3n+1
n=0 n=0
1 1 1
= tan−1 − (−1)m+1 tan−1 3m+1 → tan−1
α α α
as m → ∞.

284 VOLUME 55, NUMBER 3

 ADVANCED PROBLEMS AND SOLUTIONS

(ii) We have
!
1 1
−1 1 −1 1 −1 α2n−1 + α2n+1
tan + tan = tan 1 1
α2n−1 α2n+1 1 − α2n−1 · α2n+1
α + α−1
 
= tan−1
α2n − α−2n
√
5 1
= tan−1 2n 2n
= tan−1 ,
α −β F2n
and similarly
!
1 1
−1 1 −1 1 −1 α2n−1
− α2n+1
tan − tan = tan 1 1
α2n−1 α2n+1 1 + α2n−1 · α2n+1
α − α−1
 
= tan−1
α2n + α−2n
1 1
= tan−1 2n 2n
= tan−1 .
α +β L2n
Using the above identities, we have
m m
" 2  2 #
X
−1 1 −1 1
X
−1 1 −1 1
tan tan = tan − tan
n=1
F2n L2n n=1 α2n−1 α2n+1
1 2
2
1 2
    
1
= tan−1 − tan−1 2m+1 → tan−1
α α α
as m → ∞.
Also solved by Kenneth B. Davenport, Dmitry Fleischman, and David Terr.

Sums of Products of Binomial Coefficients

H-777 Proposed by Kiyoshi Kawazu, Izumi Kubo, and Toshio Nakata, Japan.
(Vol. 53, No. 4, November 2015)
For any nonnegative integers n, m, l prove that
n  2 X     2l
2n−2l

X n 2k 2n − 2k m−i l n−l if m = 2l;
(−1) =
k i m−i 0 if m = 2l + 1.
k=0 i≥0

Solution by the proposers.

P∞ k n
For any nonnegative integer n and formal power series f (x) = k=0 ak x , let [x ]f (x)
denote an . Let a(n, m) be the left-hand side of the identity to be proved. Then we have
n  2
m
X n
a(n, m) = [z ] (1 + z)2k (1 − z)2n−2k ,
k
k=0

since we obtain
X 2k2n − 2k 
2k
(1 + z) (1 − z) 2n−2k
= (−1)m−i z m for n≥k≥0
i m−i
m,i

AUGUST 2017 285

THE FIBONACCI QUARTERLY

using the binomial theorem. We have

a(n, m) = [z m tn ]{(1 + z)2 t2 + 2(1 + z 2 )t + (1 − z)2 }n ,
since
n   
n 2 2 n
X n n n
[t ]{(1 + z) t + (1 − z) } (1 + t) = (1 + z)2k (1 − z)2n−2k .
k n−k
k=0
Using the trinomial theorem,
 
X n
a(n, m) = [z m tn ] (1 − z)2k0 (2 + 2z 2 )k1 (1 + z)2k2 tk1 +2k2 ,
k0 , k1 , k2
k0 +k1 +k2 =n
where the sum runs over all nonnegative integers k0 , k1 , k2 satisfying k0 + k1 + k2 = n. Em-
phasizing the coefficient of tn , we have
X n

m
a(n, m) = [z ] 2n−2k (1 + z 2 )n−k (1 − z 2 )2k .
k, n − 2k, k
k

Since all terms among the sum are polynomials in z 2 , we have that a(n, m) = 0 if m = 2l + 1
for some integer l ≥ 0. So, suppose that m = 2l. Letting y = z 2 , we have
X n

l
a(n, 2l) = [y ] 2n−2k (1 + y)n−2k (1 − y)2k
k, n − 2k, k
k
X  n 2k 
= [y l ] 2n−2k (1 + y)n−2k (1 − y)2k
2k k
k
XX n   
n l n 2k
= [x y ] {2x(1 + y)} 2−2k (1 + y)−2k (1 − y)2k
n
2k k
k
{2x(1 + y)}2k
 
n l
X 2k −2k
= [x y ] 2k+1
2 (1 + y)−2k (1 − y)2k
(1 − 2x(1 + y)) k
k

1 X 2k   x(1 − y) 2k

n l
= [x y ]
1 − 2x(1 + y) k 1 − 2x(1 + y)
k
 2 !−1/2
1 x(1 − y)
= [xn y l ] 1−4
1 − 2x(1 + y) 1 − 2x(1 + y)
= [xn y l ]{(1 − 4x)(1 − 4xy)}−1/2 .
The second equality and the fourth equality hold by
zk
     X n
n n 2k
= and zn = ,
k, n − 2k, k 2k k n
k (1 − z)k+1
respectively. Since
∞   ∞  
X −1/2 X 2k k
(1 − 4x)−1/2 = (−4x)k = x ,
k k
k=0 k=0
we have
∞ X
∞   
X 2k 2l
(1 − 4x)−1/2 (1 − 4xy)−1/2 = xk+l y l .
k l
k=0 l=0

286 VOLUME 55, NUMBER 3

 ADVANCED PROBLEMS AND SOLUTIONS

Hence,   
n l −1/2 2l 2n − 2l
a(n, 2l) = [x y ]{(1 − 4x)(1 − 4xy)} = .
l n−l
Also solved by Dmitry Fleischman.

A Series with Reciprocals of Products of Fibonacci Numbers

H-778 Proposed by Hideyuki Ohtsuka, Saitama, Japan.

(Vol. 53, No. 4, November 2015)
Prove that √
∞
X 1 5−3
√ = .
n
(− 5) F2 F4 F8 · · · F2n 2
n=1

Solution by the proposer.

√
Let Fn′ = 5Fn . For n ≥ 2, we have
n
(−1)n (−1)n+1 (−1)n (α2 F2′ n + 1)
n − =
α2 F2′ F4′ · · · F2′ n−1 n+1
α2 F2′ F4′ · · · F2′ n
n+1
α2 F2′ F4′ · · · F2′ n
n n n
(−1)n (α2 (α2 − β 2 ) + 1
=
α2n+1 F2′ F4′ · · · F2′ n
n+1
(−1)n α2
= 2n+1 ′ ′
α F2 F4 · · · F2′ n
(−1)n
= ′ ′ .
F2 F4 · · · F2′ n
Using the above identity, we have
m m
X 1 X (−1)n
√ =
n=1
(− 5)n F2 F4 F8 · · · F2n F ′ F ′ F ′ · · · F2′ n
n=1 2 4 8
m 
(−1)n (−1)n+1

−1 X
= ′ + − 2n+1 ′ ′
F2 α2n F2′ F4′ · · · F2′ n−1 α F2 F4 · · · F2′ n
n=2
√
−1 1 (−1)m+1 5−3
= ′ + 4 ′ − 2m+1 ′ ′ ′
→
F2 α F2 α F2 F4 · · · F2m 2
as m → ∞.
Also solved by Kenneth B. Davenport and Dmitry Fleischman.

Late acknowledgement. Both Kenneth B. Davenport and J. M. Jarvie (solution submit-

ted via Kenneth B. Davenport) solved H-767.

AUGUST 2017 287