Documenti di Didattica
Documenti di Professioni
Documenti di Cultura
J362. Let a, b, c, d be real numbers such that abcd = 1. Prove that the following inequality holds:
ab + bc + cd + da
1
1
1
1
+ 2+ 2+ 2
2
a
b
c
d
Solution by Henry Ricardo, New York Math Circle, Tappan, NY, USA
We use the Cauchy-Schwarz inequality to see that
ab + bc + cd + da =
X
cyclic
1/2
1/2
X abcd
X 1 1
X 1
X 1
ab =
=
cd
c d
c2
d2
cyclic
cyclic
cyclic
cyclic
X 1
1
1
1
1
= 2 + 2 + 2 + 2.
=
2
c
a
b
c
d
cyclic
J364. Consider a triangle ABC with circumcircle . Let O be the center of and let D, E, F be the midpoints
of minor arcs BC, CA, AB respectively. Let DO intersect again at a point A0 . Define B 0 and C 0
similarly. Prove that
[ABC]
1.
[A0 B 0 C 0 ]
Note that [X] denotes the area of figure X.
Proposed by Taimur Khalid, Coral Academy of Science, Las Vegas, USA
R2
(sin(2A) + sin(2B) + sin(2C)) ,
2
and the conclusion follows by trivial application of Jensens inequality because the sine function is strictly
concave in (0 , 180 ), equality holding iff A = B = C, ie iff ABC is equilateral.
Case 2: If ABC is either right- or obtuse-angled, wlog at A, and D, E, F are taken as the midpoints of
arcs BC, CA, AB which do not contain A, B, C, then
1
1
DOE = DOC + COE = 180 BOC + COA = A + B,
2
2
and similarly for F OD, whereas EOF is calculated as in Case 1, yielding again
[A0 B 0 C 0 ] =
R2
(sin(A + B) + sin(B + C) + sin(C + A)) ,
2
whereas
[ABC] = [AOB] [BOC] + [COA] =
=
R2
(sin(2B) + sin(2C) sin(360 2A)) =
2
R2
(sin(2A) + sin(2B) + sin(2C)) .
2
Now, sin(2A) 0 because A is right or obtuse, while 12 sin(2B) = sin B cos B < sin B = sin(A + C), and
similarly 12 sin(2C) < sin(A + B). Applying finally Jensens inequality we obtain 21 sin(2B) + 12 sin(2C)
sin(B + C), or the proposed inequality holds strictly in this case.
Mathematical Reflections 6 (2015)
Case 3: If ABC is obtuse-angled, wlog at A, and D, E, F are taken as the midpoints of minor arcs
BC, CA, AB, or if ABC is right-angled, and D is taken as the midpoint of arc BC which contains A, then
BOD = COD = B +C, COE = B, and DOE = COD COE = C, or DF E = C2 , and similarly
DEF = B2 . Therefore,
[A0 B 0 C 0 ] = [DEF ] = 2R2 sin
B
C
B+C
B
C
A
sin sin
= 2R2 sin sin cos
2
2
2
2
2
2
2
2
> 12 .
We conclude that the proposed inequality is true, with equality iff ABC is equilateral, as long at D, E, F
are defined as the midpoints of arcs BC, CA, AB which do not contain A, B, C. It is false when ABC is
obtuse if D, E, F are defined as the midpoints of minor arcs BC, CA, AB.
Also solved by Polyahedra, Polk State College, FL, USA; Tolibjon Ismoilov, academic lyceum S.H.Sirojiddinov,
Tashkent, Uzbekistan; Nermin Hodzic, University of Tuzla at Tuzla, Bosnia and Herzegovina; Nikos Kalapodis, Patras, Greece; Adnan Ali, Student in A.E.C.S-4, Mumbai, India; Joel Schlosberg, Bayside, NY;
Alan Yan, Princeton Junction, New Jersey, United States of America; Neculai Stanciu and Titu Zvonaru,
Romania; Hyun Min Victoria Woo, Northfield Mount Hermon School, Mount Hermon, MA, USA.
J365. Let x1 , x2 , , xn be nonnegative real numbers such that x1 + x2 + + xn = 1. Find the minimum
possible value of
x1 + 1 + 2x2 + 1 + + nxn + 1.
Proposed by Nguyen Viet Hung, Hanoi University of Science, Vietnam
u + 1 + kv + 1 u + v + 1 + 1,
with equality iff v = 0.
Proof: Squaring both sides yields the equivalent inequality
p
(k 1)v + 2 (u + 1)(kv + 1) 2 u + v + 1.
Now,
(u + 1)(kv + 1) u v 1 = (ku + k 1)v 0,
with equality iff v = 0, and clearly (k 1)v 0 with equality iff v = 0. The Claim follows.
By trivial induction over n, we can now prove that
x1 + 1 + 2x2 + 1 + + nxn + 1 2 + n 1,
with equality iff x1 = 1 and x2 = x3 = = xn = 0. Indeed, the result for n = 2 is equivalent to the Claim
with k = 2. If the result is true for n 1, note that by the Claim with k = n we have
p
x1 + 1 + nxn + 1 (x1 + xn ) + 1 + 1,
with equality iff xn = 0, or renaming x1 + xn as x1 and applying the hypothesis of induction, the conclusion
follows.
Also solved by Hyun Min Victoria Woo, Northfield Mount Hermon School, Mount Hermon, MA; Adithya
Bhaskar, Mumbai, India; Neculai Stanciu and Titu Zvonaru, Romania; ngel Plaza, Department of Mathematics, University of Las Palmas de Gran Canaria, Spain; Adnan Ali, Student in A.E.C.S-4, Mumbai,
India; Paul Revenant, Lycee, Champollion, Grenoble, France; Nermin Hodzic, University of Tuzla at Tuzla,
Bosnia and Herzegovina; Polyahedra, Polk State College, FL, USA; Joel Schlosberg, Bayside, NY, USA.
s
6+
r
1.
2R
X
X
X
r
A
A
A
B
1+
sin2 + 2
sin + 2
sin sin .
2R
2
2
2
2
cyclic
cyclic
cyclic
X
cyclic
X r
X z r
yz
yz
xy
+
+
,
(z + x)(x + y)
(z + x)(x + y)
x + y (y + z)(z + x)
cyclic
cyclic
cyclic
Now
3(x + y)(y + z)(z + x)
cyclic
and
s X
X p
xy(y + z)(z + x) 3
xy(y + z)(z + x) 2(xy + yz + zx),
cyclic
cyclic
X
cyclic
xy(y + z)(z + x) =
1
(xy yz)2 + (yz zx)2 + (zx xy)2 0.
2
Senior problems
S361. Find all integers n for which there are integers a and b such that (a + bi)4 = n + 2016i.
Proposed by Titu Andreescu, University of Texas at Dallas, USA
Solution by Brian Bradie, Christopher Newport University, Newport News, VA, USA
By the arithmetic mean - geometric mean inequality,
(1 a)(1 b)(1 c)(1 d)
4abcd
4
4
,
1
1
1
+
(4 a b c d)4 .
a+b+c+d
4 108
Let x = a + b + c + d. Then we must show that
1
1
1
+
(4 x)4 ,
x
4 108
for 0 < x 4. This is equivalent to
(x 1)2 (x 4) (x 5)2 + 2 0,
which is clearly true for 0 < x 4. Equality holds for x = 1 and x = 4, which translates to a = b = c = d =
and a = b = c = d = 1, respectively.
1
4
Also solved by Hyun Min Victoria Woo, Northfield Mount Hermon School, Mount Hermon, MA; Yong
Xi Wang, Affiliated High School of Shanxi University; Adnan Ali, Student in A.E.C.S-4, Mumbai, India;
Li Zhou, Polk State College, Winter Haven, FL, USA; Jorge Ledesma, Faculty of Sciences UNAM, Mexico
City, Mexico; Daniel Lasaosa, Pamplona, Spain; Nermin Hodzic, University of Tuzla at Tuzla, Bosnia and
Herzegovina; Nicuor Zlota Traian Vuia Technical College, Focani, Romania; Paolo Perfetti, Universit
degli studi di Tor Vergata, Roma, Italy.
S363. Determine if there are distinct positive integers n1 , n2 , ..., nk1 such that
(3n21 + 4n22 + ... + (k + 1)n2k1 )3 = 2016(n31 + n32 + ... + n3k1 )2
Proposed by Titu Andreescu, University of Texas at Dallas, USA
(k + 1)(k + 2)
2
2
2025
(k + 1)(k + 2)
45 k 8
2
10
S364. Let a, b, c be nonnegative real numbers such that a 1 b c and a + b + c = 3. Prove that
a
b
c
2(a2 + b2 + c2 ) 5
+
+
+ .
b + c c + a a + b 3(ab + bc + ca) 6
Proposed by Marius Stnean, Zalu, Romnia
Solution by Hyun Min Victoria Woo, Northfield Mount Hermon School, Mount Hermon, MA, USA
Assign variables s = ab + bc + ca and p = abc. So, we have to prove the inequality
2(a + b + c)2 4s 5
3p 2(a + b + c)s + (a + b + c)3
+ ,
(a + b + c)s p
3s
6
which is equivalent to
3p 6s + 27
18 4s 5
+ .
3s p
3s
6
9s(s 2)
. Since a 1 b c, we have that (a 1)(1
5s + 12
a + b + c 1 ab bc ca + abc 0 p s 2.
If s < 2, then the result is trivially true. Otherwise, 3s = 3(ab + ac + bc) (a + b + c)2 = 9, hence
9s(s 2)
s 3 5s + 12 9s, so p s 2
.
5s + 12
Also solved by Adnan Ali, Student in A.E.C.S-4, Mumbai, India; Nermin Hodzic, University of Tuzla at
Tuzla, Bosnia and Herzegovina; Paolo Perfetti, Universit degli studi di Tor Vergata, Roma, Italy; Daniel
Lasaosa, Pamplona, Spain.
11
S365. Let
ak =
(k 2 + 1)2
,
k4 + 4
k = 1, 2, 3,
2n+1
.
n2 + 2n + 2
Solution by Brian Bradie, Christopher Newport University, Newport News, VA, USA
Let ck = k 2 + 1 and dk = k 4 + 1. Then
ck1 ck+1 =
(k 1)2 + 1 (k + 1)2 + 1
= (k 2 2k + 2)(k 2 + 2k + 2) = k 4 + 4 = dk ,
so
ak =
c2k
.
ck1 ck+1
n
Y
=
an+1k
k
k=1
2n(n1)
c1
cn0
n
Y
c2n+2k
k
cn+1k cn+1k
k+1
k=1 k1
n
Y c2n+22k
1
k
n+2k nk c
c
ck
n+1
k=2 k
cn+1
1
1
1n1
cn0
cn+1
cn+1
1
cn0 cn+1
2n+1
.
n2 + 2n + 2
Also solved by Arkady Alt, San Jose, CA, USA; Hyun Min Victoria Woo, Northfield Mount Hermon
School, Mount Hermon, MA; Adithya Bhaskar, Mumbai, India; Li Zhou, Polk State College, Winter Haven,
FL, USA; Adnan Ali, Student in A.E.C.S-4, Mumbai, India; ngel Plaza, Department of Mathematics, University of Las Palmas de Gran Canaria, Spain; G. C. Greubel, Newport News, VA; Joel Schlosberg, Bayside,
NY; Jorge Ledesma, Faculty of Sciences UNAM, Mexico City, Mexico; Daniel Lasaosa, Pamplona, Spain;
Moubinool Omarjee Lyce Henri IV, Paris, France; Nermin Hodzic, University of Tuzla at Tuzla, Bosnia
and Herzegovina; Mamedov Shatlyk, School of Young Physics and Maths No. 21. Dashoguz, Turkmenistan;
Albert Stadler, Herrliberg, Switzerland; Neculai Stanciu and Titu Zvonaru, Romania; Alessandro Ventullo,
Milan, Italy; WSA, L.T Orizont, Moldova.
12
1 1 1 1
+ + +
a b
c d
2
70
.
ab + bc + cd + da + ac + bd
Proposed by Marius Stnean, Zalu, Romnia
9 + ( + + + )2
6 a b
c d
ab + ac + ad + bc + bd + cd
9+
1 2 4r3 2 70p2
27q 2 s8 + 8p2 q 2 r6 35p2 s8
p ( 4 )
6
s
6q 2
q 2 pr, r2 qs q 4 p2 r2 p2 qs q 3 p2 s q p 3 s 3
2
r2 qs p 3 s 3 r p 3 s 3
Using these two we get
4
26
16
14
27q 2 s8 + 8p2 q 2 r6 27 p 3 s 3 + 8 p 3 s 3
Hence it is suffices to prove
4
26
16
14
27 p 3 s 3 + 8 p 3 s 3 35p2 s8
Using AmGm inequality we have
4
3
27 p s
26
3
+8p
16
3
14
3
35
35
236
3
814
3
13
Undergraduate problems
U361. Consider all possible ways one can assign the numbers 1 through 10 with a nonnegative probability so
that the probabilities sum to 1. Let X be the number selected. Suppose that E[X]k = E[X k ] for a
given integer k 2. Find the number of possible ways of assigning these probabilities.
Proposed by Mehtaab Sawhney, Commack High School, New York, USA
which are clearly equal if one of the pi s is 1 and the others are 0. We show that these are the only possibilities.
2 ++9p9
If p10 = 1, then we are done. So consider that p10 < 1. Let x = pp1 +2p
. Then x 9. By the
1 +p2 ++p9
k
convexity of t for t 0, we have
E[X]k = [(1 p10 )x + p10 10]k (1 p10 )xk + p10 10k E[X k ],
where the first inequality is equality if and only if p10 = 0. Thus p10 = 0 and E[X]k = xk . Repeating
this argument, we get p9 = 1 or p9 = 0, and so on. We conclude that there are 10 ways to assign these
probabilities.
Also solved by Daniel Lasaosa, Pamplona, Spain; Albert Stadler, Herrliberg, Switzerland.
14
U362. Let
X
Sn =
q i+j+k ,
1i<j<kn
1i<j
1i<j
X
1 X i+1 q 2i+2
q6
1
3i+3
q
q
=
=
.
1q
1 q2
(1 q)(1 q 2 )
(1 q)(1 q 2 )(1 q 3 )
1i
1i
Also solved by Daniel Lasaosa, Pamplona, Spain; Shohruh Ibragimov, National University of Uzbekistan,
Tashkent, Uzbekistan; Adnan Ali, A.E.C.S-4, Mumbai, India; Li Zhou, Polk State College, FL, USA; Yong
Xi Wang, High School of Shanxi University, China.
15
U363. Let a be a positive number. Prove that there is a number = (a), 1 < < 2, such that
X
a
a
= 2 + a 1 .
j
bac + 1
j=0
i=1
= a1
bac+1
2
bac + 2i
i=1
=
1
n+2
(n + 2)(n + 3) (n + 2)(n + 3)(n + 4)
(2 x) (5 x) (2 x) (6 x)
+
+
(n + 2) (n + 5) (n + 2) (n + 6)
2x
3x
(2 x)(3 x)(4 x)
5x
= 1
1
1
< 1,
n+2
n+3
(n + 2)(n + 3)(n + 4)
n+5
since n and x cannot both be 0. Thus < 2.
1
Next, if n 2, then n+x
n+2 2 , so > 1.
Now consider n = 0 and 0 < x < 1. Then
x
x(x 2)(x 3) x(x 2) (x 5)
= 2
+
+
+
2!
4!
6!
(1 + 1)x + (1 1)x 2
2 2x
=
=
> 1,
x1
1x
where the inequality follows from 2x < 1 + x.
It remains to consider n = 1 and 0 x < 1. Then
x + 1 (x + 1)(x 2)(x 3) (x + 1)(x 2) (x 5)
= 4
+
+
+
3!
5!
7!
2 (1 + 1)x+1 (1 1)x+1 2(x + 1)
4 (1 + x 2x )
=
=
.
x(x 1)
x(1 x)
Let f (x) = 4 (1 + x 2x ) x(1 x). Then f (0) = f (1) = 0 and f 0 (x) = 3 + 2x 2x+2 ln 2. Note that
f 0 (0) = 3 4 ln 2 > 0, f 0 (1) = 5 8 ln 2 < 0, and f 0 (x) has only one 0 for 0 < x < 1. Hence, for 0 < x < 1,
f (x) > 0 and thus > 1. For x = 0, = 4(1 ln 2) > 1 by LHopitals rule.
Finally, for the second part, we start with | sin a| = sin x and use the well-known fact that sinx =
(x)(1 x). Therefore,
Z 1
Z 1
a a 1 a(a)(2 x)
a1
1x
=
=a
t (1 t) dt a
ta1 dt = 1,
sin x bac + 1 (a + 2 x)
0
0
completing the proof.
Mathematical Reflections 6 (2015)
16
U364. Evaluate
Z
5x2 x 4
dx.
x5 + x4 + 1
dx
5
4
2
x +x +1
x +x+1
Z
1
5(2x + 1) + 3
= ln |x5 + x4 + 1|
dx
2
x2 + x + 1
Z
Z
5
2x + 1
3
dx
5
4
= ln |x + x + 1|
dx
2
x2 + x + 1
2
(x + 21 )2 + 34
5
3 2
5
4
2
1 2x + 1
= ln |x + x + 1| ln |x + x + 1|
tan
2
2
3
3
5
2x
+
1
5x2 x 4
dx =
x5 + x4 + 1
17
bxc
dx =
n Z
X
k=1
bxc
dx =
k1
n
X
k1
dx =
n
X
k1
k=1
ek1 =
k=1
en 1
.
e1
(b) Let N = ben c, or ln(N ) n < ln(N + 1). Note that bex c = k for all ln(k) x < ln(k + 1), or
Z
n
x
be cdx =
0
N
1 Z ln(k+1)
X
k=1
ln(k)
be cdx +
be cdx =
ln N
N
1
X
k=1
ln(k+1)
dx + N
ln(k)
dx =
ln N
18
U366. If f : [0, 1] R is a convex and integrable function with f (0) = 0, prove that
Z
Z
f (x)dx 4
1
2
f (x)dx.
1
f (x) dx =
2
Z
f (1 x) dx
f (x) dx +
0
f (x) + f (1 x)
dx
2
0
Z 1
x + (1 x)
f
dx
2
0
1
= f
.
2
Z
1
2
1
1
= 2
(1 x)f (0) + xf
dx
2
Z0 1
1
dx
2
f (1 x) 0 + x
2
0
Z 1
2
= 4
f (x) dx,
Z
19
Olympiad problems
O361. Determine the least integer n > 2 such that there are n consecutive integers whose sum of squares is a
perfect square.
Proposed by Alessandro Ventullo, Milan, Italy
(mod 3),
(mod 4),
(mod 4),
20
O362. Let (Fn ), n 0, with F0 = 0, F1 = 1, Fn+1 = Fn + Fn1 for all n 1. Prove that the following
identities hold:
(a) FF3n
= 2 Fn1 2 + Fn+1 2 Fn1 Fn+1 .
n
F1 = 5n .
F2n3 + + 2n+1
F2n1 + 2n+1
F2n+1 + 2n+1
(b) 2n+1
n
2
1
0
Proposed by Nguyen Viet Hung, Hanoi University of Science, Vietnam
Solution by Brian Bradie, Christopher Newport University, Newport News, VA, USA
Let (Fn ) denote the Fibonacci sequence; that is, F0 = 0, F1 = 1, and Fn+1 = Fn + Fn1 for all n 1. The
Binet formula for the nth Fibonacci number is
n n
1+ 5
1 5
Fn =
, where =
and =
.
2
2
5
Note
= 1,
and
2 +
1
= 5,
1
= 5,
1
1
= 2 + 2 = 2 + 2 = 3.
2
1 2n2
+ 2n+2 + 2(1)n ,
5
so that
2
2(Fn1
2
Fn+1
)
=
=
1
2
1
8
2
2
2n
2
+ 2 +
+ 2 2n + (1)n
5
5
6 2n 6 2n 8
+ + (1)n .
5
5
5
Moreover,
Fn1 Fn+1 =
=
=
1 2n
n 1 n+1 n+1 n1 + 2n
5
1 2n
n1 n1 (2 + 2 ) + 2n
5
1 2n 1 2n 3
+ + (1)n .
5
5
5
Thus,
2
2
2(Fn1
+ Fn+1
) Fn1 Fn+1 = 2n + 2n + (1)n =
F3n
.
Fn
21
2n + 1
k
F2n+12k
n
2n+1
X
X 2n + 1
2n + 1
=
F2n+12k +
F2n+12k
k
k
k=0
k=n+1
n
n
X
X
2n + 1
2n + 1
=
F2n+12k +
F2k(2n+1)
k
2n + 1 k
k=0
k=0
n
X
2n + 1
= 2
F2n+12k ,
k
k=0
Fn = (1)
Fn
and
n
k
=
n
nk
.
Now,
2n+1
X
k=0
=
=
2n + 1
k
F2n+12k
!
2n+1
X 2n + 1
2n + 1
2n+12k
2n+12k
k
k
k=0
k=0
!
2n+1
2n+1
X 2n + 1
X
1
2n
+
1
4n+22k
4n+22k (2n+1)
(2n+1)
k
k
5
k=0
k=0
"
#
2n+1
2n+1
1
1 + 2
1 + 2
5
1 2n+1 2n+1
5
+ 5
= 2 5n .
5
1
2n+1
X
Thus,
n
2n+1
X
1 X
2n + 1
2n + 1
F2n+12k =
F2n+12k = 5n .
k
k
2
k=0
k=0
Also solved by Arkady Alt, San Jose, CA, USA; Daniel Lasaosa, Pamplona, Spain; Nermin Hodzic, University of Tuzla at Tuzla, Bosnia and Herzegovina; Jorge Ledesma, Faculty of Sciences UNAM, Mexico City,
Mexico; ngel Plaza, University of Las Palmas de Gran Canaria, Spain; Adnan Ali, A.E.C.S-4, Mumbai,
India; Li Zhou, Polk State College, FL, USA; Ravoson Vincelot, Lycee Henri IV, Paris, France.
22
xyz
xyz
= 2 xy + yz + zx +
= 2016.
3
3
23
O364. (a) If n = p1e1 pe22 pekk , where pi are distinct primes, find the value of
X n(d)
d|n
e
X
(pi )
i=0
pi
=1+
e
X
p1
i=1
1
.
=1+e 1
p
Hence,
X n(d)
d|n
k
k
Y
Y
1
ek 1
e1 1
= nf (n) = n
1 + ei 1
= p1
pk
[pi (1 + ei ) ei ] .
pi
i=1
i=1
(b) Notice that 9312 = 97 96 and x 6 0 (mod 97). So we consider x = 97q + r, with 0 q 95 and
1 r 96. Then xx r97q+r rq+r (mod 97). Now for any prime p, it is well known that if d|(p 1),
then rd 1 (mod p) has exactly d roots. (See G. H. Hardy & E. M. Wright, An Intro. to the Theory of
Numbers, 5th ed., Oxford, p. 85.) Also, recall that the order of r is the least positive integer such that rd 1
(mod p). Hence, there are exactly (d) roots r of order d. For each such r of order d, there are 96/d values
of q such that d|(q + r). Therefore, the answer to the question is the special case of (a) for n = 96 = 25 3:
X 96(d)
d|96
Also solved by Albert Stadler, Herrliberg, Switzerland; Joel Schlosberg, Bayside, NY, USA; David E.
Manes, Oneonta, NY, USA; Adnan Ali, A.E.C.S-4, Mumbai, India.
24
O365. Prove or disprove the following statement: there is a non-vanishing polynomial P (x, y, z) with integer
coefficients such that P (sin u, sin v, sin w) = 0 whenever u + v + w = 3 .
Proposed by Albert Stadler, Herrliberg, Switzerland
25
O366. In triangle ABC, let A1 , A2 be two arbitrary isotomic points on BC. We define points B1 , B2 CA
and C1 , C2 AB similarly. Let `a be the line passing through the midpoints of segments (B1 C2 ) and
(B2 C1 ). We define lines `b and `c similarly. Prove that all three of these lines are concurrent.
Proposed by Marius Stnean, Zalu, Romnia
We use barycentric coordinates and the usual Conways notations with reference to the triangle ABC.
Points A1 , B1 , C1 have the followings absolute coordinates
e
be
cf f
ad d
,
, B1
, 0,
, C1
, ,0
A1 0,
a
a
b
b
c
c
where d, e, f are parameters. Then points A2 , B2 , C2 have absolute coordinates
d ad
e
be
f cf
A2 0, ,
, B2
, 0,
, C2
,
,0
a
a
b
b
c
c
Midpoints of segments (B1 C2 ) and (B2 C1 ) and the others similarly.
MB1 C2 (bf + ce : b(c f ) : c(b e)),
MA1 C2 (af : 2ac af cd : cd),
Equations of lines la , lb , lc .
la : (ec bf )x + (2bc 3ce bf )y + (3bf + ce 2bc)z = 0
lb : (af + 3cd 2ac)x + (af cd)y + (2ac 3af cd)z = 0
lc : (2ab ae 3bd)x + (bd + 3ae 2ab)y + (bd ae)z = 0
Mathematical Reflections 6 (2015)
26
3bf + ce 2bc
2ac 3af cd = 0
bd ae
27