Professional Documents
Culture Documents
CMJ Problems and Solutions Sep2010
CMJ Problems and Solutions Sep2010
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
doi:10.4169/074683410X510335
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 as a pdf, TEX, or Word
attachment (preferred) or by mail to the address provided above. Whenever possible, a proposed prob-
lem 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 December
15, 2010.
PROBLEMS
931. Proposed by Sam Vandervelde, St. Lawrence University, Canton, NY.
Let f (x) be a polynomial having real coefficients. Show that
f ( f (x) + x) = f (x)g(x),
for some polynomial g(x). Furthermore, prove that the remainder upon dividing g(x)
by f (x) is equal to f (x) + 1.
Prove that
1 1 4
27
f (x) d x ≥
4
f (x) d x .
0 4 0
VOL. 41, NO. 4, SEPTEMBER 2010 THE COLLEGE MATHEMATICS JOURNAL 329
Let ABC be an acute-angled triangle and let H be its orthocenter. If ta , tb , tc are the
inradii of triangles HBC, HCA, and HAB, respectively, prove that
√
ta + tb + tc ≤ (6 3 − 9)r,
where r is the inradius of triangle ABC.
934. Proposed by José Luis Dı́az-Barrero, Universitat Politécnica de Catalunya,
Barcelona, Spain.
Let a1 , a2 , . . . , an be the lengths of the sides of an n-gon with semi-perimeter s. Prove
that
ai2 a 2j n−1 n
≤ (s − ak )2 .
a
1≤i< j ≤n i
2
+ a 2
j (n − 2) 2
k=1
935. Proposed by Tom Beatty, Florida Gulf Coast University, Ft. Myers, FL.
Two players play the following game. First, player B chooses an irrational number
β > 0, then player A chooses an n ∈ N. Now they reverse roles and player A chooses
an irrational number α > 0, then player B chooses an m ∈ N with the conditions that
α = β and m = n. Next, the number of elements of the set S = { jα : j ∈ N} ∪ {kβ :
k ∈ N} in the intervals [m, m + 1] and [n, n + 1] are counted. Player A is declared the
winner if there are n elements of S in [m, m + 1] but not m such elements in [n, n + 1].
Player B wins if [n, n + 1] contains m elements of S, but [m, m + 1] does not contain
n such elements. All other outcomes are considered a draw. Prove or disprove that
there is a winning strategy for either player.
SOLUTIONS
The Gamma function and an infinite product
906. Proposed by Ovidiu Furdui, University of Toledo, Toledo, OH.
Solution by William Seaman, Albright College, Reading, PA and Michael Vowe, Ther-
wil, Switzerland (independently).
k
z (−1)n−1
For each z ∈ C, let p(z, k) = 1+ . Then
n=1
n
k
1+ z k
2n(z + 2n − 1)
p(z, k) = 2n−1
= .
n=1
1 + 2nz n=1
(2n − 1)(z + 2n)
and
k
(2k)! n=1 (z + 2n)
k
(2n − 1)(z + 2n) = ,
n=1
2k k!
we can find
z
2
2k
z k! · k 2 z· n=1 (z+ n)
p(z, k) = 2 · 2 ·
z
k z · .
2 z
2
· n=1 2
+n (2k)!(2k)z
k!k z
Since (z) = lim for Re z > 0 and z = 0, −1, −2, . . . ,
k→∞ z(z + 1)(z + 2) · · · (z + k)
z 2
lim p(z, k) = 2z−2 z [(z)]−1 .
k→∞ 2
k
2m + 2n − 1 2n
(a) Since = p(2m, k),
n=1
2n − 1 2m + 2n
∞
2m + 2n − 1 2n [(m)]2
= 22m−2 · (2m) ·
n=1
2n − 1 2m + 2n (2m)
((m − 1)!)2 4m
= 22m−2 · (2m) · = 2m .
(2m − 1)! m
∞
z (−1)n−1 2z−2 z[( 2z )]2
(b) 1+ = .
n=1
n (z)
Also solved by A RKADY A LT , San Jose, CA; M ICHEL B ATAILLE , Rouen, France; A RIN C HAUDHURI, Raleigh,
NC; H ONGWEI C HEN, Christopher Newport U.; C HIP C URTIS, Missouri Southern State U.; B RUCE DAVIS, St.
Louis C.C.; JAMES D UEMMEL , Bellingham, WA; B ILL D UNN III, Lone Star C.-Montgomery; K EITH E KBLAW,
Walla Walla, WA; M ICHAEL G OLDENBERG and M ARK K APLAN (jointly), Baltimore Poly. Inst.; G.R.A.20
P ROBLEM S OLVING G ROUP, Rome, Italy; G. C. G REUBEL, Newport News, VA; J EFFREY G ROAH, Lone Star
C.-Montgomery; E UGENE H ERMAN, Grinnell C.; H OFSTRA U NIVERSITY P ROBLEM S OLVERS, Hofstra U.;
S ANTIAGO DE L UX Á (student) and Á NGEL P LAZA (jointly), U. of Las Palmas de Gran Canaria, Spain; J ERRY
M ASUDA, Wentworth Military Acad. and C.; J ERRY M INKUS, San Francisco, CA; DARRYL N ESTER, Bluffton
U.; N ORTHWESTERN U NIVERSITY M ATH P ROBLEM S OLVING G ROUP, Northwestern U.; PAOLO P ERFETTI,
Dipartimento di Matematica, Università degli studi di Tor Vergata Roma, Rome, Italy; J OHN S AYER, Livermore,
CA; J OEL S CHLOSBERG, Bayside, NY; P ETER S IMONE , U. of Nebraska Medical Center; S PECIAL T OPICS IN
M ATHEMATICS C LASS, Dana C.; J OHN S UMNER and A IDA K ADIC -G ALEB (jointly), U. of Tampa; F RANCISCO
V IAL (student), Pontificia U. Católica de Chile, Santiago, Chile; A LBERT W HITCOMB, Castle Shannon, PA; and
the proposer.
VOL. 41, NO. 4, SEPTEMBER 2010 THE COLLEGE MATHEMATICS JOURNAL 331
An infinite sum with the Pell-Lucas numbers
907. Proposed by Brian Bradie, Christopher Newport University, Newport News, VA.
For each nonnegative integer n, let
2
n
n
an = Qj − Q 2 j +1 ,
j =0 j =0
∞
an
.
n=0
n!
n
Q j = 2Pn+1 ,
j =0
and
n
Q 2 j +1 = (2Pn+1 )2 − 1 − (−1)n
j =0
∞
an ∞
1 ∞
(−1)n 1
Thus, = + =e+ .
n=0
n! n=0
n! n=0
n! e
Also solved by A RKADY A LT (2 solutions), San Jose, CA; M ICHEL B ATAILLE , Rouen, France; P HILIP
B ENJAMIN, Berkeley C., NJ; M IHAELA B LANARIU, Columbia C. Chicago; M EGAN B ODEKOR (student),
SUNY Brockport; G ASTON B ROUWER, Macon State C.; S TAN B YRD and O SSAMA S ALEH (jointly), U. of
Tennessee at Chattanooga; C HARLES C OOK and M ICHAEL B ACON (jointly), Sumter, SC; T IM C ROSS, King
Edward’s School, Birmingham, UK; C HIP C URTIS, Missouri Southern State U.; S ERGIO FALCON and Á NGEL
P LAZA (jointly), U. of Las Palmas de Gran Canaria, Spain; H ABIB FAR, Lone Star C.-Montgomery; D MITRY
F LEISCHMAN, Santa Monica, CA; M ICHAEL G OLDENBERG and M ARK K APLAN (jointly), Baltimore Poly.
Inst.; G. C. G REUBEL, Newport News, VA; E UGENE H ERMAN, Grinnell C.; H OFSTRA U NIVERSITY P ROBLEM
S OLVERS, Hofstra U.; Y OUNG H O K IM, Inst. of Sci. Ed. for the Gifted and Talented, Yonsei U., Seoul, Korea;
PANAGIOTIS K RASOPOULOS, Athens, Greece; DANIEL L ONG (student), U. of Tennessee at Chattanooga; Y OZO
M IKATA, Bechtel, Niskayuna, NY; DAVID N ACIN, William Paterson U.; R ITURAJ N ANDAN, MEMC Electronic
Materials, St. Peters, MO; DARRYL N ESTER, Bluffton U.; ROB P RATT , Raleigh, NC; M ICHAEL S ARVER, Cedar
Crest C.; J OHN S AYER, Livermore, CA; J OEL S CHLOSBERG, Bayside, NY; W ILLIAM S EAMAN, Albright C.;
J OHN S UMNER and A IDA K ADIC -G ALEB (jointly), U. of Tampa; F RANCISCO V IAL (student), Pontificia U.
Católica de Chile, Santiago, Chile; M ICHAEL VOWE , Therwil, Switzerland; L UYUAN Y U, Tianjin U., Tianjin,
China; and the proposer.
Tnk
Tn+1 = for n > 1.
exp(k n /n k )
∞ 1the sequence converges for any real, positive t < exp(ζ (k)), where ζ (k) =
Show
i=1 i k .
Solution by Michael Andreoli, Miami Dade College, Miami, FL; Michel Bataille,
Rouen, France; James Duemmel, Bellingham, WA; and Eugene Herman, Grinnell
College, Grinnell, IA (independently).
Since 0 < t < exp(ζ (k)),
kn
ln(T1 ) = ln t and ln(Tn+1 ) = k ln(Tn ) − ,
nk
for n ∈ N. An easy induction leads to
n
1
ln(Tn+1 ) = k n
ln t − ,
j =1
jk
and lim k n = ∞. It follows that lim ln(Tn+1 ) = −∞, and hence lim Tn = 0.
n→∞ n→∞ n→∞
Also solved by A RIN C HAUDHURI, Raleigh, NC; C HIP C URTIS, Missouri Southern State U.; M ICHAEL G OLD -
ENBERG and M ARK K APLAN (jointly), Baltimore Poly. Inst.; H OFSTRA U NIVERSITY P ROBLEM S OLVERS,
Hofstra U.; J EAN H UANG (student), Harvard U.; PANAGIOTIS K RASOPOULOS, Athens, Greece; K EE -WAI
L AU, Hong Kong, China; C ONGXIAO L IU and M OHAMMED K ARIM (jointly), Alabama A&M U.; DARRYL
N ESTER, Bluffton U.; PAOLO P ERFETTI, Dipartimento di Matematica, Università degli studi di Tor Vergata
Roma, Rome, Italy; J OHN S AYER, Livermore, CA; J OEL S CHLOSBERG, Bayside, NY; W ILLIAM S EAMAN,
Albright C.; S ANGHUN S ONG, Inst. of Sci. Ed. for the Gifted and Talented, Yonsei U., Seoul, Korea; J OHN
S UMNER and A IDA K ADIC -G ALEB (jointly), U. of Tampa; T ONY TAM, Calexico, CA; L UYUAN Y U, Tianjin
U., Tianjin, China; H ONGBIAO Z ENG, Fort Hays State U.; and the proposer.
Editor’s Note. Darryl Nester of Bluffton University noted that {Tn } diverges if ln t > ζ(k), and hence {Tn }
converges iff t < exp(ζ(k)).
VOL. 41, NO. 4, SEPTEMBER 2010 THE COLLEGE MATHEMATICS JOURNAL 333
Solution by Hongbiao Zeng, Fort Hays State University, Hays, KS.
Since D is the reflection of L on X ,
By (1) and (2), ∠ILX = ∠IC C, and hence ∠ALC = ∠AC I . Therefore, ALC ∼
AC I , and hence
AB AC
= .
AC AB
Thus, AB C ∼ ACB, and hence ∠AB C = ∠ACB. Therefore, BCC B is a cyclic
quadrilateral.
Also solved by J OSHUA A LSUP (student), Columbia State U.; G EORGE A POSTOLOPOULOS, Messoloughi,
Greece; H ERB B AILEY, Rose-Hulman Inst. of Tech.; M ICHEL B ATAILLE , Rouen, France; M IHAELA B LA -
NARIU , Columbia C. Chicago; C HIP C URTIS, Missouri Southern State U.; H ABIB FAR , Lone Star C.-
Montgomery; M ICHAEL F REITAS (student), SUNY Cortland; M ICHAEL G OLDENBERG and M ARK K APLAN
(jointly), Baltimore Poly. Inst.; H OFSTRA U NIVERSITY P ROBLEM S OLVERS, Hofstra U.; G EOFFREY K AN -
DALL , Hamden, CT; Y OUNG H O K IM and S ANGHUN S ONG (jointly), 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; C ONGXIAO
L IU and M OHAMMED K ARIM (jointly), Alabama A&M U.; Y OZO M IKATA, Bechtel, Niskayuna, NY; J ERRY
M INKUS, San Francisco, CA; A NUPA M URALI, Concord, NH; DARRYL N ESTER, Bluffton U.; J OS É N IETO,
U. del Zulia, Venezuela; D ORIS S CHATTSCHNEIDER, Bethlehem, PA; J OEL S CHLOSBERG, Bayside, NY; A LEX
S ONG (student), Detroit Country Day School; H UNTER S PINK (student), Western Canada High School; S USAN
S UN (student), West Vancouver Secondary School in Vancouver, British Columbia, Canada; RONALD T IBERIO,
Natick, MA; M ICHAEL VOWE , Therwil, Switzerland; C HEN WANG (student), U. of Michigan & Shanghai Jiao
Tong U.-Joint Inst., Shanghai, China; H AOHAO WANG and J ERZY W OJDYLO (jointly), Southeast Missouri State
U.; A LBERT W HITCOMB, Castle Shannon, PA; and the proposer.
Editor’s Note. Darryl Nester of Bluffton University noted that when ∠B is sufficiently large B is between B and
Z , so BCC B is no longer a simple quadrilateral.
Prove that
(m, n)
(1, 2)
is a square of an integer for every pair of integers m and n.
Solution by Eugene Herman, Grinnell College, Grinnell, IA.
More generally, we assume there exist integers p, q, r with r = ±1 such that
Let A(m, n) denote the given matrix in the problem, so that (m, n) = det(A(m, n)).
First note that
since interchanging columns 2 and 3 and interchanging rows 2 and 3 changes A(m, n)
into A(n, m). Next note that, for all n ∈ Z, there exist integers αn , βn , γn such that
an = αn a2 + βn a1 + γn a0 .
Furthermore, since we can use any three consecutive terms of {an } as our initial terms,
we have
Specifically, we use these five equations in conjunction with the column and row
operations
C3 − βk+1 C2 → C3 , R3 − βk+1 R2 → R3 ,
VOL. 41, NO. 4, SEPTEMBER 2010 THE COLLEGE MATHEMATICS JOURNAL 335
and also use the fact that the determinant is a linear function of each column and row
of the matrix:
(m, m + k)
⎛ ⎞
a0 am am+k
= ⎝
det am a2m a2m+k ⎠
am+k a2m+k a2m+2k
⎛ ⎞
a0 am αk+1 am+1 + γk+1 am−1
= ⎝
det am a2m αk+1 a2m+1 + γk+1 a2m−1 ⎠
am+k a2m+k αk+1 a2m+k+1 + γk+1 a2m+k−1
⎛ ⎞ ⎛ ⎞
a0 am am+1 a0 am am−1
= αk+1 det⎝ am a2m a2m+1 ⎠ + γk+1 det⎝ am a2m a2m−1 ⎠
am+k a2m+k a2m+k+1 am+k a2m+k a2m+k−1
⎛ ⎞
a0 am am+1
= αk+1 det⎝ am a2m a2m+1 ⎠
αk+1 am+1 + γk+1 am−1 αk+1 a2m+1 + γk+1 a2m−1 αk+1 a2m+2 + γk+1 a2m
⎛ ⎞
a0 am am−1
+ γk+1 det⎝ am a2m a2m−1 ⎠
αk+1 am+1 + γk+1 am−1 αk+1 a2m+1 + γk+1 a2m−1 αk+1 a2m + γk+1 a2m−2
⎛ ⎞ ⎛ ⎞
a0 am am+1 a0 am am+1
= αk+1 det am
2 ⎝ a2m a2m+1 + αk+1 γk+1 det am
⎠ ⎝ a2m a2m+1 ⎠
am+1 a2m+1 a2m+2 am−1 a2m−1 a2m
⎛ ⎞ ⎛ ⎞
a0 am am−1 a0 am am−1
+ γk+1 αk+1 det am ⎝ a2m a2m−1 + γk+1 det am
⎠ 2 ⎝ a2m a2m−1 ⎠
am+1 a2m+1 a2m am−1 a2m−1 a2m−2
⎛ ⎞
a0 am am−1
= αk+1 (m, m + 1) + γk+1 (m, m − 1) + 2γk+1 αk+1 det⎝ am
2 2
a2m a2m−1 ⎠.
am+1 a2m+1 a2m
(4)
It remains to show that expression (4) can be written as p (1, 2) for some integer p.
2
Let B(m) denote the matrix in (5), and note that B(1) = A(1, 2). In fact, we show
next that det(B(m)) = det(A(1, 2)) = (1, 2). If m > 1, we use instances of the re-
cursion relation (1) with the column and row operations
C3 − pC2 − qC1 → C3 , R3 − p R2 − q R1 → R3 ,
C1 → C2 → C3 → C1 , R1 → R2 → R3 → R1
as follows:
⎛ ⎞ ⎛ ⎞
a2m−2 a2m−1 ra2m−3 a2m−2 a2m−1 a2m−3
det(B(m)) = det ⎝a2m−1 a2m ra2m−2 ⎠ = r det ⎝ a2m−1 a2m a2m−2 ⎠
a2m a2m+1 ra2m−1 ra2m−3 ra2m−2 ra2m−4
⎛ ⎞ ⎛ ⎞
a2m−3 a2m−2 a2m−1 a2m−4 a2m−3 a2m−2
= r 2 det ⎝a2m−2 a2m−1 a2m ⎠ = det ⎝a2m−3 a2m−2 a2m−1 ⎠
a2m−4 a2m−3 a2m−2 a2m−2 a2m−1 a2m
= det(B(m − 1)).
(m, m + 1) = γ1−m
2
(1, 2). (6)
We evaluate the third term in (4) much as we did (m, m + 1). The column and row
operations
VOL. 41, NO. 4, SEPTEMBER 2010 THE COLLEGE MATHEMATICS JOURNAL 337
⎛ ⎞
a2m−2 a2m−1 a2m
= −r γ2−m γ1−m det ⎝a2m−1 a2m a2m+1 ⎠
a2m a2m+1 a2m+2
= −r γ2−m γ1−m det(B(m))
= −r γ2−m γ1−m (1, 2). (7)
(m, m + k) = αk+1
2
γ1−m
2
(1, 2) + γk+1
2
γ2−m
2
(1, 2) − r 2αk+1 γk+1 γ2−m γ1−m (1, 2)
= (αk+1 γ1−m − r γk+1 γ2−m )2 (1, 2).
Also solved by JAMES D UEMMEL , Bellingham, WA; Y OZO M IKATA, Bechtel, Niskayuna, NY; W ILLIAM
S EAMAN, Albright C.; J OHN S UMNER and A IDA K ADIC -G ALEB (jointly), U. of Tampa; and the proposer.
Editor’s Note. In problem 910, the proposer assumed that {an } is a tribonacci sequence defined with arbitrary
integers a0 , a1 , a2 , not all zero, and an+3 = an+2 + an+1 + an for all integers n.