Mathematical Association of America
Mathematical Association of America
Mathematical Association of America
Source: The American Mathematical Monthly, Vol. 117, No. 6 (June-July 2010), pp. 558-565
Published by: Mathematical Association of America
Stable URL: http://www.jstor.org/stable/10.4169/000298910X492862 .
Accessed: 25/03/2013 00:46
Your use of the JSTOR archive indicates your acceptance of the Terms & Conditions of Use, available at .
http://www.jstor.org/page/info/about/policies/terms.jsp
.
JSTOR is a not-for-profit service that helps scholars, researchers, and students discover, use, and build upon a wide range of
content in a trusted digital archive. We use information technology and tools to increase productivity and facilitate new forms
of scholarship. For more information about JSTOR, please contact support@jstor.org.
Mathematical Association of America is collaborating with JSTOR to digitize, preserve and extend access to
The American Mathematical Monthly.
http://www.jstor.org
PROBLEMS
11496 (April, 2010, p. 370) Correction: On the left, square s(A A T ) and s(B B T ).
11509. Proposed by William Stanford, University of Illinois-Chicago, Chicago, IL. Let
m be a positive integer. Prove that
m 2
2
−m+1 m −2m+1
1
m 2 = 2m−1 .
k−m
k=m k k m m
11513. Proposed by Pál Péter Dályay, Szeged, Hungary. For a triangle with area F,
semiperimeter s, inradius r , circumradius R, and heights h a , h b , and h c , show that
2s F 10r (5R − r )
5(h a + h b + h c ) ≥ + 18r ≥ .
Rr R
doi:10.4169/000298910X492862
558
c THE MATHEMATICAL ASSOCIATION OF AMERICA [Monthly 117
SOLUTIONS
Perpendicular Half-Area
11392 [2008, 855]. Proposed by Omran Kouba, Higher Institute for Applied Science
and Technology, Damascus, Syria. Let the consecutive vertices of a regular n-gon P
1 (2k + 1)π
Ak Bk = Ak Ck − Bk Ck = Ak Ak+1 − ρ sin φ − +π
2 n
π
(2k + 1)π
= sin + ρ sin φ − ,
n n
(2k + 1)π
Bk M = Ck O + Hk M = Ck O + ρ cos φ − +π
n
π
(2k + 1)π
= cos − ρ cos φ − .
n n
Therefore,
2 Area (M Ak Bk ) = Ak Bk · Bk M
sin(nβ/2) 1
sin(α + kβ) = sin α + (n − 1)β .
k=0
sin(β/2) 2
560
c THE MATHEMATICAL ASSOCIATION OF AMERICA [Monthly 117
n−1
2(k + 1)π n−1
2(2k + 1)π
sin φ− =0= sin 2φ − .
k=0
n k=0
n
Summing both sides of (1) over k, we obtain the required result:
n−1
n 2π
2 Area (M Ak Bk ) = sin ,
k=0
2 n
and this last expression gives the area of P.
Also solved by M. Bataille (France), D. Beckwith, R. Chapman (U.K.), C. Curtis, J. Freeman, D. Grinberg, J.-P.
Grivaux (France), K. Hanes, E. A. Herman, S. Hitotumatu (Japan), E. J. Ionascu, L. R. King, P. T. Krasopoulos
(Greece), J. H. Lindsey II, O. P. Lossers (Netherlands), V. Mihai (Canada), C. R. Pranesachar (India), M. A.
Prasad (India), R. A. Russell, A. Stadler (Switzerland), R. Stong, M. Tetiva (Romania), J. Vinuesa (Spain),
A. Vorobyov, Z. Vörös (Hungary), M. Vowe (Switzerland), GCHQ Problem Solving Group (U.K.), Microsoft
Research Problems Group, and the proposer.
Concurrent Lines
11393 [2008, 856]. Proposed by Cosmin Pohoata, student, National College “Tudor
Vianu,” Bucharest, Romania. In triangle ABC, let M and Q be points on segment
AB, and similarly let N and R be points on AC, and P and S, points on BC. Let
d1 be the line through M, N , d2 the line through P, Q, and d3 the line through
R, S. Let ρ(X, Y, Z) denote the ratio of the length of XZ to that of XY. Let m =
ρ(M, A, B), n = ρ(N , A, C), p = ρ(P, B, C), q = ρ(Q, B, A), r = ρ(R, C, A),
and s = ρ(S, C, B). Prove that the lines (d1 , d2 , d3 ) are concurrent if and only if
mpr + nqs + mq + nr + ps = 1.
Solution by Michel Bataille, Rouen, France. We use barycentric coordinates rela-
tive to (A, B, C), and accordingly we write U (u 1 , u 2 , u 3 ) as an abbreviation for
“U = (u 1 A + u 2 B + u 3 C)/(u 1 + u 2 + u 3 ).” (When u 1 + u 2 + u 3 = 0 we obtain a
“point at infinity”). With this convention we have M(m, 1, 0), N (n, 0, 1), P(0, p, 1),
Q(1, q, 0), R(1, 0, r ), and S(0, 1, s). The equation of line d1 is
x m n
y 1 0
= 0, that is, x = my + nz.
z 0 1
Jensenoid Inequalities
11399 [2008, 948]. Proposed by Biaggi Ricceri, University of Catania, Catania, Italy.
Let (, F , μ) be a measure space with finite nonzero measure M, and let p > 0.
Let f be a lower semicontinuous function on R with the property that f has no global
minimum, but for each λ > 0, the function t → f (t) + λ|t| p does have a unique global
minimum. Show that exactly one of the two following assertions holds:
(a) For every u ∈ L p () that is not essentially constant,
1/ p
1
Mf |u(x)| dμ
p
< f (u(x)) dμ,
M
Proof. Let λ be positive and let λn be a sequence of positive numbers converging
to λ. Letting tn = φ(λn ), we have f (t) + λn |t| ≥ f (tn ) + λn |tn |. Let λ0 be such that
0 < λ0 < λ and m = infR ( f (t) + λ0 |t|). Now
f (tn ) + λn |tn | = f (tn ) + λ0 |tn | + (λn − λ0 )|tn | ≥ m + (λn − λ0 )|tn |.
This proves that for all t, (λn − λ0 )|tn | ≤ f (t) + λn |t| − m, so that for n large enough
that λn − λ0 > 12 (λ − λ0 ), taking t = 0 gives |tn | < 2( f (0) − m)/(λ − λ0 ). Thus tn
is bounded. Let t be a limit point of tn . There exists a subsequence tψ(n) which
converges to t . For all t in R, f (t) + λn |t| ≥ f (tn ) + λn |tn |. By lower semicontinuity,
for all t,
f (t) + λ|t| = lim inf f (t) + λψ(n) |t| ≥ lim inf f (tψ(n) ) + λψ(n)
tψ(n)
= lim inf f (tψ(n) ) + λ|t | ≥ f (t ) + λ|t |.
562
c THE MATHEMATICAL ASSOCIATION OF AMERICA [Monthly 117
Proof. Let λn be a sequence such that limn→∞ λn = +∞, and let tn = φ(λn ). For
t ∈ R, we have f (tn )/λn + |tn | ≤ f (t)/λn + |t|, and in particular f (tn )/λn + |tn | ≤
f (0)/λn . Let λ0 be a fixed positive value, and let m = infR [ f (t) + λ0 |t|]. Now f (tn ) ≥
m − λ0 |tn |, so (1 − λ0 /λn )|tn | ≤ f (0) − m /λn . Therefore limn→∞ tn = 0.
For the other claim of the lemma, let λn be a positive sequence that tends to zero,
let tn = φ(λn ), and let t be a limit point of tn (if one exists). The argument of Lemma
1 proves that for any real t, f (t) ≥ f (t ). That makes f (t ) a global minimum for f ,
contrary to the hypothesis. Since tn has no limit point, limn→∞ |tn | = +∞.
From these two lemmas, we see that the range of φ contains (0, ∞) or (−∞, 0)
(but not both). We will show that in the first case conclusion (a) holds. Similarly, the
second case leads to (b).
Assume
the range
contains (0, ∞), and let m(λ) = inf R f (t) + λ|t| . Now f (t) ≥
supλ m(λ) − λ|t| . If t = φ(λ), then f (φ(λ)) = m(λ) − λ|φ(λ)|. Thus f is the point-
wise supremum of a family of affine functions on (0, ∞), so f is convex there. We
claim that f is actually strictly convex. Indeed, if f is affine on some interval [a, b]
with 0 < a < b, then we can choose λ such that the function f λ given by f λ (t) =
f (t) + λ|t| reaches its infimum at a point of (a, b). Since f λ is is affine on this in-
terval, it is minimized at an interior point only if it is constant on that interval, which
contradicts the uniqueness of the minimum point.
Let s, t be given with t > 0 and −t ≤ s < t. There exists λ such that t = φ(λ).
Thus
f (s) + λ|s| > f (t) + λ|t| ≥ f (t) + λ|s|.
We obtain f (s) > f (t). (If −t ≤ s ≤ t, we obtain f (s) ≥ f (t).)
For the integral
inequality, we have −|u(x)| ≤ u(x) ≤ |u(x)|. So f (u(x)) ≥ f |u(x)| . Since f is
convex, Jensen’s inequality yields
f (u) ≥ f |u| ≥ f |u| .
Squares On Graphs
11402 [2008, 949]. Proposed by Doru Catalin Barboianu, Infarom Publishing,
Craiova, Romania Let f : [0, 1] → [0, ∞) be a continuous function such that
f (0) = f (1) = 0 and f (x) > 0 for 0 < x < 1. Show that there exists a square
with two vertices in the interval (0,1) on the x-axis and the other two vertices on the
graph of f .
Solution by Byron Schmuland and Peter Hooper, University of Alberta, Edmonton, AB,
Canada. Extend f by letting f (x) = 0 for x ≥ 1. Define g(x) = f (x + f (x)) − f (x)
for x ≥ 0. If there exists x ∈ (0, 1) with g(x) = 0, then a square as required exists with
vertices
(x, 0), (x + f (x), 0), (x, f (x)), (x + f (x), f (x)).
This is derived by writing sin2 nφ in terms of exponentials and summing the resulting
geometric series. Now divide (i) by r and integrate with respect to r to get
∞
rn 1 (1 − r )2 + 4r sin2 φ
sin nφ = log
2
. (ii)
n=1
n 4 (1 − r )2
Differentiate (i) with respect to r to obtain
∞
1 1 (r − 1)2 − 2(r 2 + 1) sin2 φ
nr n−1 sin2 nφ = − . (iii)
n=1
2(1 − r )2 2 [(1 − r )2 + 4r sin2 φ]2
The limit at r = −1 in (ii) gives us
∞
(−1)n−1 1
sin2 nφ = − log cos φ.
n=1
n 2
Now we can write the requested limit as
∞
−2 rn sin2 nx
lim x lim 1 − 2 2 sin2 nφ.
x→0 r →−1+ n n x
n=1
564
c THE MATHEMATICAL ASSOCIATION OF AMERICA [Monthly 117
A Minimum Determinant
11415 [2009, 180]. Proposed by Finbarr Holland, University College Cork, Cork, Ire-
land. Let (A1 , . . . , An ) be a list of n positive-definite 2 × 2 matrices of complex num-
bers. Let G be the group of all unitary 2 × 2 complex matrices, and define the function
F on the Cartesian product G n by
n
∗
F(U ) = F(U1 , . . . , Un ) = det Uk Ak Uk .
k=1
Show that
n
n
minn F(U ) = σ1 (Ak ) · σ2 (Ak ),
U ∈G
k=1 k=1