Problems and Solutions

Source: The College Mathematics Journal, Vol. 44, No. 3 (May 2013), pp. 233-240
Published by: Mathematical Association of America
Stable URL: http://www.jstor.org/stable/10.4169/college.math.j.44.3.233 .
Accessed: 11/04/2013 07:02

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 College Mathematics Journal.

http://www.jstor.org

This content downloaded from 134.68.189.135 on Thu, 11 Apr 2013 07:02:41 AM

All use subject to JSTOR Terms and Conditions
PROBLEMS AND SOLUTIONS
EDITORS
Curtis Cooper Shing S. So
CMJ Problems CMJ Solutions
Department of Mathematics and Department of Mathematics and
Computer Science Computer Science
University of Central Missouri University of Central Missouri
Warrensburg MO 64093 Warrensburg MO 64093
email: cooper@ucmo.edu email: so@ucmo.edu

This section contains problems intended to challenge students and teachers of college mathematics.
We urge you to participate actively BOTH by submitting solutions and by proposing problems that are
new and interesting. To promote variety, the editors welcome problem proposals that span the entire
undergraduate curriculum.
Proposed problems should be sent to Curtis Cooper, either by email (preferred) as a pdf, TEX, or
Word attachment or by mail to the address provided above. Whenever possible, a proposed problem
should be accompanied by a solution, appropriate references, and any other material that would be
helpful to the editors. Proposers should submit problems only if the proposed problem is not under
consideration by another journal.
Solutions to the problems in this issue should be sent to Shing So, either by email as a pdf,
TEX, or Word attachment (preferred) or by mail to the address provided above, no later than August 15,
2013.

PROBLEMS
1001. Proposed by Michel Bataille, Rouen, France.
Given real numbers a, b such that 0 < a < b and m = a+b 2
< π4 , evaluate
Z m− Z b 
lim f (x) d x + f (x) d x ,
→0+ a m+

where
(1 + cos(2m − 2x)) cos(a − x) cos(b − x)
f (x) = .
(1 − sin(a − x))(1 − sin(b − x)) sin(2m − 2x)
1002. Proposed by Mowaffaq Hajja, Yarmouk University, Irbid, Jordan.
Let ABCD be a convex quadrilateral. Prove that there exists a point P inside ABCD
such that [PAB] = [PBC] = [PCD] = [PDA], if and only if the diagonals bisect each
other. Here [ABC] denotes the area of triangle ABC.
1003. Proposed by D. M. Bătinetu-Giurgiu, Matei Basarab National College, Bucha-
rest, Romania and Neculai Stanciu, George Emil Palade Secondary School, Buzău,
Romania.
Prove that in any triangle ABC
1 1 1 27
+ 4 + 4 ≥ 4,
wa
4 wb wc s
where s is the semiperimeter of ABC and wa is the length of the internal bisector of
angle A opposite side a.
http://dx.doi.org/10.4169/college.math.j.44.3.233

VOL. 44, NO. 3, MAY 2013 THE COLLEGE MATHEMATICS JOURNAL 233

This content downloaded from 134.68.189.135 on Thu, 11 Apr 2013 07:02:41 AM

All use subject to JSTOR Terms and Conditions
1004. Proposed by Ángel Plaza and Sergio Falcón, Universidad de Las Palmas de
Gran Canaria, Spain.
Let m be a nonnegative integer and
m+1  
k(k+1)/2 m + 1
X
Um = (−1) .
k=0
k

Prove that for n, a nonnegative integer,

(a) U4n = 0,
(b) U4n+2 = (−4)n+1 , and
(c) U2n+1 = (−1)(n+1)(n+2)/2 2n+1 .

1005. Proposed by Elias Lampakis, Kiparissia, Greece.

Prove that for any triangle ABC
 
X ra − r 4R + r
ra ≥ ,
ra + r 2
where r is the inradius of the triangle, R is the circumradius, and ra is the radius of the
excircle at side a. The sum is over all cyclic permutations of (ra , rb , rc ).

SOLUTIONS
An inequality with a triangle
976. Proposed by D. M. Bătinetu-Giurgiu, Matei Basarab National College, Bucha-
rest, Romania and Neculai Stanciu, George Emil Palade Secondary School, Buzău,
Romania.
For any triangle, let s, r , and R denote its semiperimeter, inradius, and circumradius.
Show that
p
(2x + y)s 2 + r (y − 2x)(4R + r ) ≥ 9 3 4x y 2 s 2 R 2r 2 ,
for any x, y > 0.
Solution by Reza Akhlaghi, Big Sandy Community and Technical College, Prestons-
burg KY; Minh Can, University of Redlands, Redlands CA; and Marian Dincǎ, Bucha-
rest, Romania (independently).
Since s 2 − 4Rr − r 2 = 21 (a 2 + b2 + c2 ), s 2 + 4Rr + r 2 = ab + bc + ca, and
16s 2r 2 R 2 = a 2 b2 c2 , we have
2x(s 2 − 4Rr − r 2 ) + y(s 2 + 4Rr + r 2 )
= x(a 2 + b2 + c2 ) + y(ab + bc + ca)
   
ab + bc + ca ab + bc + ca
= x(a + b + c ) + y
2 2 2
+y
2 2
r r
3 6 6 6 6 2 2 2 2
9 x y a b c 3 xy a b c
p
≥9 6
=9 = 9 3 4x y 2 s 2 R 2r 2 .
2 4
The desired inequality follows.

234 „ THE MATHEMATICAL ASSOCIATION OF AMERICA

This content downloaded from 134.68.189.135 on Thu, 11 Apr 2013 07:02:41 AM

All use subject to JSTOR Terms and Conditions
Also solved by A RKADY A LT, San Jose CA; G EORGE A POSTOLOPOULOS, Messolonghi, Greece; M ICHEL
BATAILLE, Rouen, France; H ABIB FAR, LSCS Montgomery; D MITRY F LEISCHMAN, Santa Monica CA; S ANG -
WOO J O , Inst. of Sci. Ed. for the Gifted and Talented, Yonsei U., Seoul, Korea; E LIAS L AMPAKIS, Kiparissia,
Greece; K EE -WAI L AU, Hong Kong, China; J OSEPH L OVERDE, Foothill High School; PAOLO P ERFETTI,
Dipartimento di Matematica, Università degli Studi di Roma “Tor Vergata”; A LBERT W HITCOMB, Castle
Shannon PA; DAVID YANG; and the proposers.

Functions and countable sets

977. Proposed by Greg Oman, University of Colorado at Colorado Springs, Colorado
Springs CO.
Let X be a set and f : X → X be a function. A subset Y of X is said to be closed
under f provided that whenever y ∈ Y , f (y) ∈ Y . Prove or disprove: There exists an
uncountable set X and a function f : X → X with the following property:

for any subsets Y and Z of X which are closed under f , either Y ⊆ Z or Z ⊆ Y. (∗)

Solution by Greg Ronsse, Eastern Illinois University, Charleston IL.

Suppose that there exist an uncountable set X and a function f : X → X satisfying
(∗). For each x ∈ X , let I (x) = {x} ∪ { f n (x) : n ∈ N where f n denotes f composed
with itself n times}. Then I (x) is closed under f . Thus, it follows from (∗) that for all
x, y ∈ X , x ∈ I (y) or y ∈ I (x). Pick y1 ∈ X . Since I (y1 ) is countable, there exists
a y2 ∈ X \ I (y1 ), and hence I (y1 ) ⊂ I (y2 ). Proceeding by induction, we obtain a
sequence,

I (y1 ) ⊂ I (y2 ) ⊂ I (y3 ) ⊂ · · · ,

where the inclusions are strict, with yn+1 ∈ X \ I (yn ). Now, let
∞
[
Y = I (yn ).
n=1

Then Y is a countable set which is closed under f . Let z ∈ X \ Y . Then Y ⊂ I (z), and
hence y1 = f N (z) for some N ∈ N. It follows that

I (z) = {z, f (z), . . . , f N −1 (z)} ∪ I (y1 ).

Thus, I (z) cannot possibly contain the infinite subset {y1 , y2 , . . . } of Y , which contra-
dicts Y ⊂ I (z).
Also solved by G EORGE A POSTOLOPOULOS, Messolonghi, Greece; PAUL B UDNEY, Sunderland MA;
M ICHELLE DAHER, Notre Dame U., Zouk Mosbeh, Lebanon; JAMES D UEMMEL, Bellingham WA; D MITRY
F LEISCHMAN, Santa Monica CA; J ERROLD G ROSSMAN, Oakland U.; E UGENE H ERMAN, Grinnell C.; E UGEN
I ONASCU, Columbus State U.; L AURA I OSIP (student), U. of Edinburgh, UK; J OS É N IETO, U. del Zulia, Mara-
caibo, Venezuela; N ORTHWESTERN U NIVERSITY M ATH P ROBLEM S OLVING G ROUP; F RANCISCO P ERDOMO
and Á NGEL P LAZA (jointly), U. deLas Palmas de Gran Canaria, Spain; S TEPHEN S CHEINBERG; W ILLIAM
S EAMAN, Albright C.; J OHN S TEELMAN, Indiana PA; T IMOTHY W OODCOCK, Stonehill C.; and the proposer.

A limit with the harmonic numbers

978. Proposed by Michel Bataille, Rouen, France.

VOL. 44, NO. 3, MAY 2013 THE COLLEGE MATHEMATICS JOURNAL 235

This content downloaded from 134.68.189.135 on Thu, 11 Apr 2013 07:02:41 AM

All use subject to JSTOR Terms and Conditions
 1 1
Let Hn = 1 + 2
+ ··· + n
be the nth harmonic number. Evaluate

m
!!!
X Hk+n k+n
lim lim lim (−1)k x .
n→∞ x→1− m→∞
k=1
Hn

Solution by Eugene Herman, Grinnell College, Grinnel IA; and Haohao Wang and
Jerzy Wojdylo, Southeast Missouri State Univeristy, Cape Girardeau MO (indepen-
dently).
We use the generating function
∞
X log(1 − x)
Hj x j = − where − 1 < x < 1,
j=1
1−x

as follows:
m
X n+m
X
(−1)k Hk+n x k+n = (−1)n H j (−x) j
k=1 j=n+1
 
Xn+m n
X
= (−1)n  H j (−x) j − H j (−x) j  ,
j=1 j=1

and so
!!  
m n
X log(1 + x) X
lim lim (−1)k Hk+n x k+n = (−1)n lim − − H j (−x) j 
x→1− m→∞
k=1
x→1− 1+x j=1
 
n
log 2 X
= (−1)n+1  + H j (−1) j  .
2 j=1

We simplify the last term inside the parentheses:

n j
n X n n n
X X (−1) j X 1X X (−1)i 1 − (−1)n+1−i
H j (−1) j = = (−1) j = .
j=1 j=1 i=1
i i=1
i j=i i=1
i 2

Hence, if n is even,

n n/2
X X 1 1
H j (−1) = j
= Hn/2 ,
j=1 j=1
2j 2

and, if n is odd,

n (n+1)/2 n+1 (n+1)/2

X X 1 X 1 X 1 1
H j (−1) j = − =− + = −Hn+1 + H(n+1)/2 .
j=1 j=1
2j − 1 k=1
k k=1
2k 2

236 „ THE MATHEMATICAL ASSOCIATION OF AMERICA

This content downloaded from 134.68.189.135 on Thu, 11 Apr 2013 07:02:41 AM

All use subject to JSTOR Terms and Conditions
Therefore, since limn→∞ (Hn − γ − log n) = 0, we conclude, for n even,
m
!!!
X
k Hk+n k+n 1 Hn/2
lim lim lim (−1) x = − lim
n→∞ x→1− m→∞
k=1
Hn 2 n→∞ Hn
1 γ + log n − log 2
=− lim
2 n→∞ γ + log n
1
=− ,
2
and, for n odd,
m
!!!
X Hk+n k+n
lim lim lim (−1)k x
n→∞ x→1− m→∞
k=1
Hn

−Hn+1 + 12 H(n+1)/2
= lim
n→∞ Hn
−γ − log(n + 1) + 12 (γ + log(n + 1) − log 2) 1
= lim =− .
n→∞ γ + log n 2
Also solved by G EORGE A POSTOLOPOULOS, Messolonghi, Greece; C ON A MORE P ROBLEM G ROUP, U. of
Aarhus, Copenhagen, Denmark; JAMES D UEMMEL, Bellingham WA; D MITRY F LEISCHMAN, Santa Monica
CA; M ICHAEL G OLDENBERG, Baltimore Poly. Inst. and M ARK K APLAN, Towson U. (jointly); PAOLO P ER -
FETTI , Dipartimento di Matematica, Università degli Studi di Roma “Tor Vergata”; and the proposer.

The probability for an acute chordal triangle

979. Proposed by Steven Finch, Harvard University, Cambridge, MA.
Let A, B, C be three independently selected, uniformly distributed points on the unit
sphere S 3 in R4 . With probability one, the points A, B, C determine a unique chordal
triangle T inscribed within S 3 , with sides as straight lines through the interior of S 3 .
What is the probability that T is acute?
Solution by Northwestern University Math Problem Solving Group, Northwestern Uni-
versity, Evanston, IL.
We solve this problem generalized to the unit n-sphere. (A similar solution to this
problem appeared in The Mathematical Intelligencer, 8 (2) 1986, p. 49).
Given that the events 6 ABC ≥ π2 , 6 BCA ≥ π2 , and 6 CAB ≥ π2 are disjoint, if each
has probability Q n , then the desired probability Pn that ABC is acute will be Pn =
1 − 3Q n . Next, we find the value of Q n .
Because of the symmetry of the problem, we may assume that A is fixed. After
picking B, consider the hyperplane through B orthogonal to the line AB. This hyper-
plane divides the n-sphere into two hyper-spherical caps SIn and SIIn , with A ∈ SIIn . Then
the angle at B will be greater than π2 if C ∈ SIn , and this happens with probability equal
to the quotient between the (hyper) surface areas of SIn and S n .
Recall that the surface area of an n-sphere of radius is
2π (n+1)/2r n
An (r ) =

,
0 n+1 2
R∞
where 0 is Euler’s gamma function 0(x) = 0 e−t t x−1 dt.

VOL. 44, NO. 3, MAY 2013 THE COLLEGE MATHEMATICS JOURNAL 237

This content downloaded from 134.68.189.135 on Thu, 11 Apr 2013 07:02:41 AM

All use subject to JSTOR Terms and Conditions
 n (r, φ) of a hyper-spherical cap of an n-sphere of radius r and
The surface area Acap
colatitude angle φ can be obtained by integrating the surface area of an (n − 1)-sphere
of radius r sin α with arc element r dα over a great circle arc:

φ φ
2π n/2r n
Z Z
n (r, φ) =
Acap An−1 (r sin α)r dα = sinn−1 α dα.
0 0( n2 ) 0

−
→ −
→
If O is the center of Sn and θ is the angle between the vectors OA and OB, then the
π −θ
colatitude of SI is φ = 2 , and the probability that C ∈ SI is
n n

π −θ

Z π−θ  Z π−θ
n (1, 2 )
Acap 0 n+1 n 1 −1

2 2
= n √
2
sin α dα = B
n−1
, sinn−1 α dα,
An (1) 0 2 π 0


2 2 0

where B(x, y) = 0(x)0(y)

0(x+y)
is the beta function.
Next, we find Q n by integrating with respect to θ . Since the probability density of
θ is

n (1, θ)
−1
Acap
  
d n 1
=B , sinn−1 θ,
dθ An (1) 2 2
π−θ
!
 −2 Z π Z
n 1 2
Qn = B , sin n−1
θ sin n−1
α dα dθ.
2 2 0 0

The value of Q n can be computed for specific n. Since Pn = 1 − 3Q n , we obtain

thereby these values: P1 = 41 , P2 = 12 , P3 = 14 + π42 , P4 = 53
70
, P5 = 14 + 45π
256
2 , P6 =
1751 1 4
2002
, etc. In particular, the answer to this problem is P3 = 4 + π 2 .
Also solved by E UGEN I ONASCU, Columbus State U.; R ANDY S CHWARTZ, Schoolcraft C.; J OHN S TEELMAN,
Indiana PA; and the proposer.

Minimum area of a triangle and its angles

980. Proposed by George Apostolopoulos, Messolonghi, Greece.
Let ABCD be a square with sides of length a. Suppose K and L are points on the sides
BC and CD, respectively, so that the perimeter of triangle KCL is 2a. If triangle AKL
has minimum area, determine the measures of its angles.
Solution by Randy Schwartz, Schoolcraft College, Livonia MI.
We consider the following generalization of this problem: Let ABCD be a square
with sides of length a. Suppose K and L are points on the sides BC √ and CD, respec-
tively, so that the perimeter of triangle KCL is p where 0 < p < (2 + 2)a. If triangle
AKL has minimum area, determine the measures of its angles.
Without loss of generality, let A = (0, 0), K = (x, a), and L = (a, y) with 0 <
x < a and 0 < y < a. Then

1 −→ − → 1
area of 4AKL = kAK × ALk = (a 2 − x y).
2 2

238 „ THE MATHEMATICAL ASSOCIATION OF AMERICA

This content downloaded from 134.68.189.135 on Thu, 11 Apr 2013 07:02:41 AM

All use subject to JSTOR Terms and Conditions
Since the perimeter of 4AKL is p, we have
p
(a − x) + (a − y) + (a − x)2 + (a − y)2 = p. (1)

If p = a, (1) becomes 2x y = a 2 , and hence

a2
area of 4AKL = ,
4
for all allowable choices of points K and L.
If p 6= a, let
1 2
F(x, y, λ) = (a − x y) + λ[(x + y + p − 2a)2 + 2a(x + y) − (x 2 + y 2 + 2a 2 )].
2
Then
1
Fx (x, y, λ) = − y + 2λ[(x + y + p − 2a) + a − x] = 0, (2)
2
1
Fy (x, y, λ) = − x + 2λ[(x + y + p − 2a) + a − y] = 0, (3)
2
and

Fλ (x, y, λ) = (x + y + p − 2a)2 + 2a(x + y) − (x 2 + y 2 + 2a 2 ) = 0. (4)

Combining (2) and (3), we have

1
2λ(x − y) = (x − y).
2
If x 6= y, then λ = 41 . Substituting this into (2) or (3) leads to p = a, a case we dis-
pensed with earlier. If x = y, equation (4) becomes
 
1
x 2 + 2( p − a)x + a 2 − 2 pa + p 2 ,
2

and hence
√ !
2
x = a + p −1 + .
2

Using this, some straightforward calculations show that the minimal triangle satisfies
√ √
  
1 3
,


Area of 4AKL = p p 2− −a 2−2
2 2
2a √
 
6 AKL = 6 ALK = tan−1 2+2 −1 ,


p

and
π h p √ i
− 2 tan−1 2−2 +1 .


6 KAL =
2 2a

VOL. 44, NO. 3, MAY 2013 THE COLLEGE MATHEMATICS JOURNAL 239

This content downloaded from 134.68.189.135 on Thu, 11 Apr 2013 07:02:41 AM

All use subject to JSTOR Terms and Conditions
Also solved by R EZA A KHLAGHI, Big Sandy CTC; M ICHAEL A NDREOLI, Miami-Dade C.; H ERB BAILEY,
Rose-Hulman Inst. of Tech.; M ARK BAILEY; T HOMAS BASS, Carson-Newman C.; M ICHEL BATAILLE, Rouen,
France; E LIZABETH B ROWN, Waukesha Tech. C.; J OACHIM B ULLACHER, Quinebaug Valley C.C.; J OHN
C HRISTOPHER, California State U.; M ARGARET C IBES, Hartford CT; C HIP C URTIS, Missouri Southern State
U.; M ICHELLE DAHER and A JAJ TARABAY (jointly), Notre Dame U., Zouk Mosbeh, Lebanon; L UZ D E A LBA,
Drake U.; PAUL D EIERMANN, Southeast Missouri State U.; M ARIAN D INC Ǎ, Bucarest, Romania; ROB
D OWNES, Mountain Lakes High School; JAMES D UEMMEL, Bellingham WA; W ILLIAM D UNHAM, Muhlenberg
C.; H ANS E VANS, SFSU; J OHN F ERDINANDS, Calvin C.; D MITRY F LEISCHMAN, Santa Monica CA; M ICHAEL
G OLDENBERG, Baltimore Poly. Inst. and M ARK K APLAN, Towson U. (jointly); R AYMOND G REENWELL,
Hofstra U.; E UGENE H ERMAN, Grinnell C.; C LIFF H OLROYD, Santa Clara CA; E UGEN I ONASCU, Columbus
State U.; L AURA I OSIP (student), U. of Edinburgh, UK; PAUL I RWIN, Randolph C.; S TEPHEN K ACZKOWSKI,
South Carolina Governor’s School for Sci and Math; PANAGIOTIS K RASOPOULOS, Athens, Greece; Z ACHARY
K UDLAK, Mount St Mary C.; W EI -K AI L AI and J OSHUA S MOAK (student, jointly), U. of South Carolina
Salkehatchie; E LIAS L AMPAKIS, Kiparissia, Greece; K EE -WAI L AU, Hong Kong, China; J OSEPH L AZZARA,
Lewis Center OH; W EIPING L I, Walsh U.; N ICOLAS L OBOTSKY, Hunter C.; J ODY L OCKHART, U.S. Naval
Academy; J OSEPH L OVERDE, Foothill High School; M ASAO M ABUCHI, Hachinohe N.C.T., Japan; D OUGLAS
M AGOMO, Edison State C.; B OB M ALLSON, Indiana Wesleyan U,; D IANA M C G INNIS, Georgia Perimeter
C.; J ERRY M INKUS, San Francisco CA; Á NGEL P LAZA and F RANCISCO P ERDOMO (jointly), U. de Las
Palmas de Gran Canaria, Spain; K EVIN ROPER, Cedarville U.; W ILLIAM S EAMAN, Albright C.; E ARL S MITH,
Sun City Center FL; RON S MITH, Edison State C.; J UYEOP S ONG, Taerang Middle School, Seoul, Korea;
M IHAI S TOENESCU, Bischwiller, France; N ORA T HORNBER, Raritan Valley C.C.; M ICHAEL VOWE, Therwil,
Switzerland; H AOHAO WANG and J ERZY W OJDYLO (jointly), Southeast Missouri State U.; Q IAN WANG,
U. of Michigan-Shanghai Jiao Tong U. Joint Inst.; S TUART W ITT, Brooklyn NY; M ICHAEL W OLTERMANN,
Washington and Jefferson C.; and the proposer.

Editor’s Note. The Skidmore College Problem Group (2 solutions) was inadvertently
omitted from the list of solvers of CMJ Problem 961.

240 „ THE MATHEMATICAL ASSOCIATION OF AMERICA

This content downloaded from 134.68.189.135 on Thu, 11 Apr 2013 07:02:41 AM

All use subject to JSTOR Terms and Conditions