Mathematical 7 Olympiad ; wif ' Glance RESTO ed gee Titu Andreeccu & Razvan Geles LO earns tee Birkhauser ooTitu Andreescu yr My & 2006 | Razvan Gelca MATHEMATICAL OLYMPIAD. CHALLENGES Foreword by Mark Saul Birkhauser Boston ® Basel © BerlinTitu And a Ireescu Razvan Gelca University of Texas partm Univer at Dallas De i School of Natural Sciences and Mathematics Texas Tech Univoane sues and Statistics 2601 N. Floyd Rd., FN33 Richardson, TX 75080 Us penned USA. “ I at ing-i Wary of Congress Cataloging-in-Publication Data Auheescu, Titu, 1956- Matheinatical Otympiad challenges / Ti non ical Olympiad challenges / Titu Andreescu, Rizvan Gelea. prcludes bibliographical references. SH3N ()-8176-4155-6 (alk.) — ISBN 3-7643-4155-6 (alk.) !. Mathematics—Probl ci i onarese esd ems, exercises, etc. L Gelca, Razvan. 1967- Hl. Title. StO°.76-de2I 99-086229 CIP eters Set Cen 20; WANE OORT OS OER RCSL cs Subject Classifications 2000: 00405. Q0A07. 05-XX. O8-XX, 11-XX, §1-XX ISHN i} 0-8176. 5. 1] ser Bostoy - 4155-6 Birkha diuser B. 1] Istin ts 97%-0-8176-4 155-9 Birkhauser Boston ant 0 Birkhauser Boston ome Birkhauser, Boston, second printing ©2001 Birkhiuser Boston, third printing Or jatkhiuser Boston, fourth printing nung irkhduser Boston. fifth printing “C05 Birkhauser Boston, sixth printing ® Birkhauser hp ding , electronic adaptati imi imilar methodology now known or hereafter developed is forbidden monwareserby similar The Use j 3 e 1 Use in this publicati ve oti sntifi Pi vcation or wade Hames, trademarks, Service marks and similar terms, e en if they ai ide fied as such, is not to be taken as an expression of opinion as to whether or not they are subject Ci {0 proprictary rights. Printed in the United States of America. (TXQUHP) 9 876 SPIN 11419761] Www birkhauser.com Contents Foreword Mark Saul " Preface “pate xiii Titu Andreescu-and Razvan Gelca y I: Problems > 1 1 Geometry and Trigonometry 3 1.1 A Property of Equilateral Triangles... .-- +--+ +--+ 4 1.2 Cyclic Quadrilaterals .....---- Cee ee ee ee 6 1.3 PowerofaPoint......--- 2-2 ee tee ees 9 1.4 Dissections of Polygonal Surfaces ...-..- See eee 13 1.5 Regular Polygons .. 2... - ee ee ee tes -eees 16 1.6 Geometric Constructions and Transformations ......--.- 20 1.7. Problems with Physical Flavor ...--.-- ++ -++ eee: 22 1.8 Tetrahedra Inscribed in Parallelepipeds...... - wee ete 624 4.9 _ Telescopic Sums and Products in Trigonometry ....--- » 26 4.10 Trigonometric Substitutions ©... .-- +6 +522 e eee 29° 2 Algebra and Analysis 33 2.1 No Square is Negative ...-+--+--+:- See 34 2.2 Look at the Endpoints. 2... 2-2-2 eee tees . 36 2.3. Telescopic Sums and Products in Algebra... . +--+ + 38 2.4 Onan Algebraic Identity ...... Se ee ee eee 41. 2.5 Systems of Equations. ....-..-- eee ee 43 2.6 Periodicity. ....----+- Lee eee 47 2.7. The Abel Summation Formula .....-----+--085 . 50 28 X+1/x 2.0. ee ee eee eee eee wee ee ee 982 2.9 Matrices... 26. eee ee eee Lecce eee eee ee 54 2.10 The Mean Value Theorem .....----++-+- tee eee 5S. Matematicd, maternaticd, matematicd, atita matematicd? ‘ Nu, mai mult. : Grigore Moisil 1 Mathematics, | | mathematics, mathematics, that much mathematics? No, even more.Foreword Why Olympiads? Working mathematicians often tell us that results in the field are achieved after long experience and a deep familiarity with mathematical objects, that progress is made slowly and collectively, and that flashes of inspiration are mere punctuation in periods of sustained effort. The Olympiad environment, in contrast, demands a relatively brief period of intense concentration, asks for quick insights on specific occasions, and re- quires a concentrated but isolated effort. Yet we have found that participants in mathematics Olympiads have often gone on to become first class mathemati- cians or scientists, and have attached great significance to their early Olympiad experiences. For many of these people, the Olympiad problem is an introduction, a glimpse into the world of mathematics not afforded by the usual classroom situation. A good Olympiad problem will capture in miniature the process of creating mathematics. It’s all there: the period of immersion in the situation, the quiet examination of possible approaches, the pursuit of various paths to so- lution. There is the fruitless dead end, as well as the path that ends abruptly but offers new perspectives, leading eventually to the discovery of a better route. Perhaps most obviously, grappling with a good problem provides practice in dealing with the frustration of working at material which refuses to yield. If the solver is lucky, there will be the moment of insight which heralds the start of a successful solution. Like a well-crafted work of fiction, a good Olympiad problem tells a story of mathematical creativity which captures a good part of the real experience, and leaves the participant wanting still more. And the present book gives us more. It weaves together Olympiad prob- lems with a common theme, so that insights become techniques, tricks become methods, and methods build to mastery. While each individual problem may be a mere appetizer, the table is set here for more satisfying fare, which will take the reader deeper into mathematics than might any single problem or contest. The book is organized for learning. Each section treats a particular tech- nique or topic. Introductory results or problems are provided with solutions, then related problems are presented, with solutions in another section.uae Foreword The craft i : ognizing mites tion Olympiad coach or teacher consists largely in rec- tant skill that the coach wn Problems, Indeed, this is the single most impor- Olympiad coaches have fered to the student. In this book two master ence. Teachers will find oa | the results of their experience to a wider audi- own exercise. Olympiad stare es and topics for advanced students, or for their stronger and more ready to ok will find practice material which will leave them mathematics it may or yy n on the next challenge, from whatever corner of _ introduction to the experience eweomers to Olympiads will find an organized There is al: i in mathematics ee here for the more general reader who is interested eye, and working thre viene the problems, letting their beauty catch the standing. The mult; le € authors solutions will add to the reader’s under- not apparently related The ullons link together areas of mathematics that are a geometric transforma ey often illustrate how a simple mathematical tool— way, stretched or resha ‘0; OF an algebraic identity—can be used in a novel , ped to provide an unexpected solution to a daunti problem. o a daunting These prob! : chall engine one. Tha daunting on any level. True to its title, the book is a ementary solutions The no elementary problems—although there are el- usual high school carrie | content of the book begins just at the edge of the leaned on, either for solution The calculus is sometimes referred to, but rarely ces, standard in Euro ton or for motivation. Properties of vectors and matri- be prepared to be stymied Chena are drawn upon freely. Any reader should way of effort and patience but het Much is demanded of the reader by In this, it is not unlike inathematins cs ho is greatly repaid. x Mark Saul Bronxville School Preface At the beginning of the twenty-first century, elementary mathematics is under- going two major changes. The first one is in teaching, where one moves away from routine exercises and memorized algorithms toward creative solutions to unconventional problems. The second consists in spreading problem solving culture throughout the world. Mathematical Olympiad Challenges reflects both trends. It gathers essay-type, non-routine, open-ended problems in undergrad- uate mathematics from around the world. As Paul Halmos said, “problems are the heart of mathematics,” so we should “emphasize them more and more in the classroom, in seminars, and in the books and articles we write, to train our students to be better problem-posers and problem-solvers than we are.” The problems we selected are definitely. not exercises. Our definition of an exercise is that you look at it and you know immediately how to complete it. It is just a question of doing the work, whereas by a problem, we mean a more intricate question for which at first one has probably no clue to how to approach it, but by perseverance and inspired effort one can transform it into a sequence of exercises. We have chosen mainly Olympiad problems, because they are beautiful, interesting, fun to solve, and they best reflect mathematical ingenuity and elegant arguments. Mathematics competitions have a long-standing tradition. More than 100 years ago, Hungary and Romania instituted their first national competitions in mathematics. The Eétvés Contest in elementary mathematics has been open to Hungarian students in their last years of high school since 1894. The Gazeta Matematica contest, named after the major Romanian mathematics journal for high school students, was founded in 1895. Other countries soon followed, and by 1938 as many as 12 countries were regularly organizing national mathemat- ics contests. In 1959, Romania had the initiative to host the first International Mathematical Olympiad (IMO). Only seven European countries participated. Since then, the number has grown to more than 80 countries, from all conti- nents. The United States joined the IMO in 1974. Its greatest success came in 1994, when all six USA team members won a gold medal with perfect scores, a unique performance in the 40-year history of the IMO. Within the United States there are several national mathematical compe-xiv Preface titions, such as the AMC +> 8 (formerly the American Junior High School Mathematics Examination), AMC +> 10 (the American Mathematics Contest for students in grades 10 or below) and AMC +> 12 (formerly the Ameri- can High School Mathematics Examination), the American Invitational Math- ematics Examination (AIME), the United States Mathematical Olympiad (US- AMO), the W. L. Putnam Mathematical Competition, and a number of regional contests such as the American Regions Mathematics League (ARML). Every year more than 600,000 students take part in these competitions. The prob- lems from this book are of the type that usually appear in the AIME, USAMO, IMO and the W. L. Putnam competition, and similar contests from other coun- tries, such as Austria, Bulgaria, Canada, China, France, Germany, Hungary, India, Ireland, Israel, Poland, Romania, Russia, South Korea, Ukraine, United Kingdom, and Vietnam. Also included are problems from international com- petitions such as the: IMO, Balkan Mathematical Olympiad, Ibero-American Mathematical Olympiad, Asian-Pacific Mathematical Olympiad, Austrian-Po- lish Mathematical Competition, Tournament of the Towns, and selective ques- tions from problem books and from the following journals: Kvant (Quantum), Revista Matematica din Timisoara (T imigoara Mathematics Review), Gazeta Matematica (Mathematics Gazette), Matematika v Skole (Mathematics in School), American Mathematical Monthly, and Matematika Sofia. Over 60 problems were created by the authors and have yet to be circulated. Mathematical Olympiad Challenges is written as a textbook to be used in advanced prebiem solving courses, or as a reference source for people inter- ested in tackling challenging mathematical problems. The problems are clus- tered in thirty sections, grouped in three chapters: Geometry and Trigonometry, Algebra and Analysis, and Number Theory and Combinatorics. The placement of geometry at the beginning of the book is unusual but not accidental. The reason behind this choice is well reflected in the words of V. 1. Amot’d: “Our brain has two halves: one is responsible for the multiplication of polynomials and languages, and the other half is responsible for orientating figures in space and all things important in real life. Mathematics is geometry when you have to use both halves.” (Notices of the AMS, April 1997). Each section is self-contained, independent of the others, and focuses on one main idea. All sections start with a short essay discussing basic facts, and one Or more representative examples. This sets the tone for the whole unit. Next, a number of carefully chosen problems are listed, to be solved by the reader. The solutions to al] problems are given in detail in the second part of the book. After each solution we provide the source of the problem, if known. Even if successful in approaching a problem, the reader is advised to study the solution given at the end of the book. As problems are generally listed in increasing order of difficulty, solutions to initial problems might suggest XV Preface illuminating ideas for completing later ones. At the very end ve include a glossary of definitions and fundamental properties used in the boo ; vn classes Mathematical Olympiad Challenges has been successfully tested inc’ ase taught by the authors at the Illinois Mathematics and Science Academy. & ne University of Michigan, the University of Iowa, and in the training of t SA International Mathematical Olympiad Team. In the end, we would li © ex press our thanks to Gheorghe Eckstein, Chetan Balwe, Mircea Grecu, umi g Feng, Zvezdelina Stankova-Frenkel and Florin Pop for their Supgestions ant : especially to Svetoslav Savchev for carefully reading the manuscript an contributions that improved many solutions in the book. Titu Andreescu, American Mathematics Competitions Razvan Gelca, University of MichiganMATHEMATICAL OLYMPIAD CHALLENGES-PartI- ProblemsAxyourouosiAy, pue A1yaul0ay J tajdeyq) |; |“gy ore ayy UO St HEId + Vd = Od pure ‘Oy ue ay) UO SI Od + Vd = Ad ‘Jaq ose 24) UO SI d JI Id + Td = Ve “osed SII} Ul “aJOITOUIMAID ay) UO Std Jt Ajuo pue Jy reauspjoo are ,J pue J ‘¥ 1eYI SMOTIOY I] “ASIMIZTIO 09 pue *q pue y uaamjaq SI gf Hf OZ S! Dd V7 afaue ayy Jeroyepnba st D.dd apdueiy ayy pue jd dq 2ulf ay) UO St Y 20UIg -2]3ueL ay) 0} JO1IOIUS JOU St way) ‘aseo ayy St siqy Jf “(Z'T'T 2un31y 9as) reauyjoo ae yp pue Og ‘g syurod ay) uaym Apaureu ‘ayerauasap st pd gf 2[8ueuy ayy uaym auiu93ap sn 9] ‘Id pue ‘gd ‘¥_ oy jyenba se Vdd a[8uety ayy JO sapts oY) 1eY? SMO]IOJ YT “Dd = jd souay {yesajepinba SI D,dd 2]8uet al “Osiy “Fd = Vd OS ‘uONe}OI ay YBnoM g_ Jo 93eu OUI SE Pf Apts ay) Jey) Vas 9M ‘Wf ff 2[Buvuy ayy ye Suryoo] “py 01 sayejor g IY) ION “MONEJO! sty) YSnoiy) gf pue y Jo soBeur ayy oq ,f PUE ,Y 177 : ‘= Od pue ‘€ = gd ‘X = Vd teu gons g quiod B pue OgyY 28uel) esole] -inba ue yonssuos ‘z pue ‘A ‘x sipZuay apis WIM ZAY s[sueyy} B UdAIH *f ~ ‘aysueuny 1st 8 JO SyISugy apts oy) ae Dg pue ‘gd ‘Vd ky Ayedoud ayy yum sued ay] ul g syutod ayy Jo snooy ayy pul -afBuew persjepinba ue 9q Dey IT “9 saysueiy) Jesaqeyinba ayy Jo yySuay apis ay) puly “¢ = Od pue ‘p= gd “E = Vd yey) yons gy 2[Suery pesojyepinbs ue apisur g yuiod & sjsixa 919, 1) Zz “Gg <4 +¥d ¢ 7eU] AAOIG ‘oueyd sy ur qurod & g pue 071 = V7 YM snquioyl & aq GOGV 191 + [CVT an3iy c . < (Qd+ddt+ vay = (Dd ‘Fd ‘Va }xew Aysuyes yey DgV a[suewy pesrsyepenba ue yo sued ay} ut yf squrod ayy Jo snd0j ayi puly “¢ ‘Id < Fd + Ve ‘afsueiy ayy Jo sOLajuT ayy ut g yurod Aue 10} yey anol ‘opis jsaBuoj oy SI TY ‘OGV asuew) ul ‘Z , ‘Jerayepnba st gy ual ‘Od pure ‘gd “Wd Jenbo sapts Suraey a[sueiy & JNISUOS ULD BUO DY 3[Suely & JO 1OLayUE ous UE yf yulod Arana Joy FI yey ApsuIEU GUaWAyBIs asIZAUOD 34} aA0Ig “(LTT sin31y 29s) 3 punore astmyaojo 09 Jo ay8ue ue Aq apSueLsy oy) ayeyor sn yoy ‘spoy Auadoid sty} Aysn 998 OF, ‘suajqoid pojepar SULMOT{OJ tf} SASIo1EXa SB BARI] AAA “BOP! \uasayyip & sasn ‘1aaamoy ‘Jooid ayy, “suosAod sepnsas j[e O} papuayxa aq uvo Ausadoud SILL . ‘omy dayio ayy fo uns ayy oj ponba SI SyISUa] aasyt asayt fo auo uaYy QpoAQuINIAID ay uO Say d {I “Dd pun ‘Gd ‘Yd 0} jonba syisuay apts fo aj8uviay v jonsisuoa uvo auo ‘jg y fo ajo4iguno4ng TUT onary 2Y] UO al] jou Saop wy) q itiod v puv Dg Yy ajfuvia josawjmba uv uaary “ynuaazauiu 94) Ul ZHUIIIS Aq JOU Aimyuso yjUsa]YySta ayy Ut JapN| Aq foyytou pareacasip sem yt ‘A[Zursudimg ‘Anowioed auejd yeorssepo to ped ‘joey o[dusis @ st wos0ay) s,nraduiog “g¢6T UT niedwog “q URloNeWsyeW URI -uewoy ayy Aq paoyou sem yorya ‘sapBueny esoyepmba ayy Jo Ayrodoid Burmoy -[OF 9U) St INSEE sty] JO BoUaNbasuos y ‘Jelazepmbea sig gy s[Fuewy oy} usyy *, julod & 07.09 Ysnosy y punowe g sajej0x aUO FI ‘g pue Y sjutod omy UaAID safsuely pesgepnby yo A4yzsdorg VOT ¥ sajSueny jerwynby yo Auadoig ¥ TT ANIAUOUOSIL], pue Ayasuroan ‘| sa}deyD ¢Jo SuondastaiU! ap 2q J 197 “snOSopeue Afa}a{du09 are saseo Jaylo ayy Soundy ay) Ur Jeyy ¥T] St UOHeINSyUOS ay; a1a4M ase oY UT Jood YI OP [HEA aM. ETT andy We ajouto ay) 07 aSuer st Hef wy) aaoad sn raj pue (7D auiy ay YIM gy Jo1oWNIP Jo 9JsI9 ay JO VONDasIOUT ay) Jy Xq a1OUDP Og “GD UO sary Od A>uDSue) 94) Je) S} ssand pood ke (¢'Z" | 21NS1.4 aas) ‘umeIp Uaaq sey andy oy) JAyV “gy sasauwvip fo a]9419 aYl 01 MASUD} S1 OH I WYI AAO “| PUD J IW Sajo41d OMI aYI O} SIUaSUDI ayt ojo g fo suonsafoad ay aq © puv g 1a Tq puv 2 Ww sajoar ays styasaanut V y8nosys Surssed auyy y 5212419 omy fo sputod uounuod ays aq q PUD Y 2T : “(Kayamoary ut sula{qodd) alsIaUIOaD ap Pulay “Goi 4009 §,e2195t], 2yB1094H ur paseadde wajqosd paioajas puosas aL : “smoj]oy uorsnjouios ayy pur ‘rej us ae SOd pure Hyd sojsuern oyy wey) saydun sp “OSd7 = yOd7 ww TTT omndty syesayepupend 2194 ZT Smoys juouindae seus y “SOd7 = OXd7 wy smofoy yf yenba are sgq7 pue Ovd?7 sojdue poquosut Kg “Sg@d7 = SOd7 puooas ayy ut pue ‘Ovd7 = Ovd7 [esaepupend sry ayy uy (7-71 aNB1y 99s) sa[Sue wyBu ds0ddo omy sey Way} Jo YoRa souls ‘DAD ae S_Od pue OV sfesarefupenb oy], ‘Sd Ud = Od 2004 ‘Sd/Od = Od/ad Ajdusr tM sey eps are SO q pue Oy d sojSuery oyp rey nord ppm 3K. ‘Sd Pun Ud fo uvaut mjauoas ays st Of IYI adord “g uv Y 10 aj2419 BYE 01 sjuasuv} ays o1u0 g fo suoyrafosd ayy § puv y puv gy uo q fo uoysalosd ayt aq © 377 ‘9]2419 ays uo mod v g pup ajo419 v ut ploy? D 2g GY T “ajdurexa ue st a1951 “satadoid omy asayy Sursn Aq paajos aq ues UasoYp aaey aa suIatqord YL TTT andy “Soue OWeS oY] UL paquosur are sapsue ay) asnesaq poy saniyenba ayy suonentis yioq uy *(q°Z"[ amNL{ 2as) atsoddo st Jo juawayddns ay 0} Jenba st aj8ue ue pue ‘euoSerp Joyo ay) pur apis asoddo ayn Aq pautsoy afsue ay) 0} Jenba st feuoFeip ouo pue apis e Aq pauoy afsue ayy :eaey spesayeptpenb 4949 yey) saysodoid omy Jo asnesaq ‘spesazeytspenb d4J949 J0y JURY 0} SE SuONENHS Yons Ut apt Pood W “sa[Sue auios Jo Ajyenba aq Buraoad 0} soonpar uayo Anauioad auejd ut sura[qord uonnaduias Surajog syerayepLapend aps ZT “Base JURISUCD Sey Do pur ‘T_ “Ye Yim paunoy o[sueLn omy YoIM JOF dV a8uery jerdyepinba ue yo aueyd ayy ut g syutod ay Jo snd0j ay} PUL “6 “dV = WZ + WAt WX II a01g ‘oO2T = XWZ7 = ZWA7 = AWX7 WM Ypns souayur sy uy yuIod & W pure ‘gd = XZ pue “Wd = ZA ‘Od = AX Wm oven ay ‘7 AX Jopysuoy JOLOTUT sy UE WWod e Y puk a[SueLN [eLOyeTINbo Ue 9q DEY IT °S Anjauouo8tyy pue Ayou0an *{ sodey . 9SI Gd + Vd wnpoid ayy ‘afono oy) MUM J YBnomp Surssed ouy Arenique ue Jo g pure y SuORDasiayUt oq) Japisuoo pur ‘auerd ay) Ut a[oz19 & pue gq julod e x1y WUlOg BJOIIMOT CT “reautpjoo ore Zy pur ‘ly “Og yey) aaorg “OpSy pue EyZy Jo uoNoasioqur ayy %g pur ‘Syty pue Zyl y Jo uonsasioqur ay ty “by ty pue TyOy jo uonsasiaqul ay} aq OY 19] pue UoSexay 1949 v aq Sprytyzyl yey I] “EL “yurod paxy @ y8nomy) sassed 7g aut] ay) ‘SOUeA 2 Udy JY) VAOIg “J IE J sassaUt J OF Y WO jUaTuL) puddas auL “€E yum Koua8uen Jo wrod ay @ YHA ‘CO pur xO OF waduE) st 7 afoul ayqeisea W AEX a[Bue amp) Jo XE apis ay} uO yWFOd paxy aq WIA] “ZL spesarepinbo stg ,N,AW afduery ayy yeys aaorg “Apaatioadsal ,q pur‘, «We DWH puooas aur Joy FY pue ‘VD ‘Dg saul] ay SpesIOL INW a[Suern ayy Jo oostownouto ayy, ‘Apoatjoodsox ‘gy pur ‘yD ‘pg OO J Jo suonoofoad ay) 9q gf pur ‘NWT 'WLO7 = DLE7 = €Lv7 eH yons Jovajur sy ut jurod e aq yf Jaf pue ‘aySuri anoe ue aq DF 397 “IL “WN 01 Jenorpuadiad st Vy Wey) aA01g *kyaanoadsax “77 pue Dg swuawiBos our Jo stuodprut ay) 9q A PLR HY 19] ‘G Woy juarayip ore fy pue 7 pue “7 9 0} ejnoipuadiad si yy “Fg 0} sejnorpuadsad st Jy yey) Yons G YyBnosy Surssed aur] e uo aq 4 pur FT “Vy WO opmynye oy Jo yoog ay) J] pue a[SueIN e aq DEV 197 “OL : “7.2 uO sat] @1V Jo sguasuinono ayy yeijt arosg “FqUadUt By) 2q [191 DAY Suen ays uy “6 “OL = FL vey aaorg © 71k yaSIaqUT ,.D pue ,g_ s1owosiq ay pue .09 = V7 ‘IgV ajsurin ayl uy *g °EQ 7 = IVA Vey aA0ld "FCAT = AVA7 pure AGI] = AVA7 wy Usouy sta] “(gO J WEY JES0p9 J YIM) TIGV [exayepupenb xoeauoo ev Jo Dg opis ay) UO UAAIS ae y pue 7 SIUIOd : : ‘20+ 0d =d4 yey oaolg “Apaanoadsar ‘pg aul ayy 0) D puke y Wo sxejnoipuad: -rod ay Jo 1923 ay) O ‘ed pue ‘Dy qwaules our ayy uo ywod B 9q WW 17] “9]oISTWas B Jo yurodprut ay} y pue spurodpua ay) 9q D pur g 177 * “onpoAouoo are Yq pure ‘qa ‘og ‘gy 01 wWadsar WIM gq JO SUOHDAYa OY) WEY} BAO “UONIaSIA;UT 194) 2q _ Jot pure ‘HeINIIpuediad are TG | pue Oy sjeuoSerp oy) yey? Yons yerayepspenb xaauoo ev aq TIGV PT qutog egosamog £1 “FeAUI|JOO Be 7 pue F7 JO sioyasiq a[Sue JoMAUE puke 101Na)xa ay) OO Dg Y eSueLN ay) Jo y xazJan Jo suoNsafosd inoy oy) yey) BAOIg “p “samy Jay ayy Aq pauoy afSuetN a1 Jo raWUaD0yUO aI st Ay PUR «Zz ¢ yiGe siuyod ay Jo auo yey aaosg “Afaanoadsor ‘Yq pur ‘C.D ‘Og ‘GY Soult ay) OO J Jo suoNsaford ay aq 4 pu ‘Z ‘4 ‘x J] -xaUaA Aue wy jua “LOY IP ‘2[ufouInas9 syy UO yutod & aq ¢f 19] pue aSuRJD9I& 29g TITY 191 “€ “d47 pur d¥Q7 so[Bue ays Jo wns ayy purty *O8T = Id A7 + dd V7 Wu) Yons Gg Y a[8uepax ap Ur Uasoys st g IIod JoN|UI UY @ “Kea N pure py uaym arenbs ay) Jo 19)429 ay} Jo sno] ay) puly “g pur oO swutod amp sareredas Ay J’ yey) Yons arenbs & 9q Od NW 19 pue ‘gO Afaanoads 24 “VO Saul-sey oy) UO syutod fy pue py ‘ajBue YB e 9q FOV? IFT ‘T -saruadoid Pan Cf *3)D119 Oyy 0} UadUBI SET Os ‘GOT IG wy) saydun syL “Oy oO} Jarfesed st 7q souay SyBu st ¥Og7 alsue auL “SVTGO ‘OY ox ay Jo wiodprus amp sig OUTS “9f419 Vy) JO 1}UEd ay 2q © I>] ‘aN S| SUP AYA 20s OF “(EET 91N Sty 99S) ajorto ay) 0} JUaTUE) st ga yeyy Burjou Aq yoosd ayy wes a, : “Aq wauBas auy oy SI29S1G 2] J WY! BAdI “2I24IINU_AS BY) YN FY aur ay) fo uoMIas4anus ay? J &G pup Dg auy ay ojo q nuod ays fo uousaload ayy J dq aiouag “py 240 ay fo * -pnodpnu ay aq q 13] pun gy sa1owp fo aj24ujoqwas v uo mud v 3g 2 12T “smoys ajduzexa xou ayy se ‘surajqoid Sutajos ut [nyosn 3q wes uJ0d v fo samod Jo uoKoU a4, “jazmauos are sajoito Jo saved 2am oy Jo soxe jeoipes ayy ‘puanbasuod “FaqWad [eoIpes oyy pay[ed st puw ‘safo119 2am ay) 0} Jadsaa YIM siamod Jenba sey jurod uoRSasIa}Ut Jay. “assayuT sed puodas BY} JO JEU) PUB SafoID JO ared ISI OY} JO SIXE JeoIpeI oY} ‘s1a;UW9d JeaUT|[OOUOU YIM SefosIo dary) U2AID u qutog eyo 1amod “ET TET amsty *sjurod uorsasiaqut 4194] SUTEIUOD AysnorAago sixe jeoIpEs oy) YOOsIOIUE SafoNd OMY 9Y) JT “WOHLIIT aaoge aly AysHeS $19}U99 BY) 0} S2OUEISIP soy 7C1C uO H Jurod ayy YSnom Burssed 7QIC oj [eUOZOYVO au ay} SI SND] ay) JDUI_Y ty - w = {00 = 100 saydun 70H pue 'OgG sojsuem o7 parjdde waioayy ueaoseyyhg aup uatp (Z'¢T aInty 295) 7919 oJWO g Jo UorDafoid ay 9q 01D as00ND 9M JL ‘fa ww = fod -jod . yey SMOT[O} “CH —7Qd)(Y+2O d) SI [9419 puosas oy} 0) J9adsai IM _ JO JaMod oup ‘ApEyTUITS “(a — lod + 'Od) St 21ND IsIy ay 0 Dadsa1 YIM _ Jo Jomod om YEU aas am IQ aur ayy Buryey Aq ‘gq yurod auwios 30q *sajoI19 OM) ay} Jo Mpes ayy Zy pue ly pue siajuao 343 2q 7Q pue IC 39] ‘psopuy “szx JooIpys DY} Poy[ed st AUT] SITY “S9fo419 ay) JO S1a]UAD ayy Aq pauTUZa}ap aUO 24) 0} Je[NOIpuadiad auty B SE S9]D119 Om) 0} Jadsax yim Samod jenba Sutaey syutod ayy Jo snd0] ay, Ver aunaiy (VET anBty 098) Ad «Vd = Id “Vd soUdH{ TEPUNS 3q P[nom g,F7q pur YY Sofsuemn aIp wary ‘oor amp UO, pur ,V qua ‘g YSnonp Suyssed aur Joyoue oram ,£7,¥ Jf OOUIS “ZY UIT BH JO go1oyd ayy JO Wapuadapur sty] “afoI9 otp 07 oadsar IM g_ JO IaMod ay payed Anawouo8uuy pue Ayouoon “pseu or‘aonfuns pouo8éjod puozas ayy ou sazayd asayy Suyquassy uayy puv sazaid kuow Karas ona ssarf ays Sup &q sayi0 aig) o1t auo pauuofsuvs aq uvd vaiD aUDS ays ys saonfins joUOshjod omy kuy Te'L omsry “any st afoul yonur oy uy ‘ozenbs v 198 0} saad says a1quiassvar pue saaard Auews Ajouuy oul sozenbs oan Aue ano ued auo wy SmoYs U1aIOdY] URBOTeYIY Up Jo ([-p [ andy) Jooud pearydesd Surmojpoy 2y,L saaeying jeuosdjog Jo suonoassiq: pT “rejnaipuedied are Gy pue yyy Wey) MOYS ~y 3B aul puodas: Basil FOC pue DOV sofsuesy Jo sapououmsio yf “DW > GW pue py >-gsy iey) Kem e yons ul @ pur 2 Je afoufofusas ay} pue jy 1° AV SSN BUT] Y “GY LAWL pue GC J9}U9d JO afouTOIUNaS & JaPISUOD “ZT “yuaunguos are 4X pue ‘AG “WY sour ay) Wey) BAOIg “AV PUL g IE 7g J9j9wWeIP YEN |[d1I9 OY] S}DastoIUy dg uly ay pur ‘yy pue Dw DY saowieIp YUM ayoutD ay sSasiaqul JD oUy SUL “Z URYyIJayIO AX aut ayi uo yutod w aq g 19] “Ze Dg sIRAW AX Ul] OWL “A pue X Je asia Gg pue DY SioJUIEIP Ysa $9949 OUL “Fopso yey) UF ‘aUTT e UO syuTOd yoUNSIP snoy aq G pue ‘D ‘g ‘VT “II “aul IySteNs B UO aI OH pue‘d ‘WV sturod ayy yey) WA0Id “FAd Pue NG _ SepSuewy Jo saposrownoyd ay) JO quid uonsasiazut puosas oy aq O 12] “Ajaanoadsar ‘gD pue J g soury ayy WA A Jo suonoasiayut ayy oq H pue yf rf pue FOV ajsuen o} JOVAUt qutod Aue aq g OT “Dg 01 alesed st FC Jey) Yons ‘Ajoanoadsal ‘OV pue gy sapis ay) uo syutod aq 7 pue gj 19] pue afsuery ve 2q DEV 1 ‘ol el ssonping reuo8Ajog jo suoyassig ‘¢'l “Hf wouides oury oy) Jo yrodprus ayy st DJ Atuo pur jt 77. wawas aur ay) Jo wrodprun otp st 7 yyy DAOIY * HY 18 GY sult ay pue 4 ye S$ dforlorwas syoasiojUt 77 UIE YL “7 Ie Og pue @y saul] ay) pue 7 Ie Pasta! 7g pue Dy soul] oy] “gy J9j}ouIL Ip JO § 9portorwas e uf paquosur jexavpupenb xaauoo v aq GIEV PL 6 “‘reauryjoo axe FY saUaD0yN ay) puR ‘G ‘g wy) aorg “GO pue g sjurod ye yoassayut WO pue Ng Sia}auteIp YIM Sapouio ayy, “Apaanoadsai ‘py pue gy sapis ayy uo uayE) are AY pue yy siuiod ayy “o[8uery aoe uk aq OEY 197 “g “OW/WY one: ay) ‘Joord ym ‘putz "gy vo qutod e ye WDassaJUL 7D pue FP Jo sioyasiq se[noipuadiod ayy rey) Kem v Yons ur J pue 7 3 7 sjoosio Y YBnom Bussed aur V “AV Jo urodpiu ayy aq (7 12] pur “1D pue FY Jo LoNDasIoIUT Jo WwIOd puooas ay) aq 2) 197 “(Z9 3 g) 7% 0} qy qosuey ayy mesp [9 Uo Yjuod B wor “19 Jo soaquL amp Ut 7D YAS ‘sa]D419 d1n}UIdUOD 9q 7 pue 1939] +7 jutod uous Aue oAvy wW9y) Jo Om} OL 10 qurod uowusod & y8nosy) ssed 7 7 pue ‘Gg ‘Dy SIo}aUleIp YUM saps ay) yO JeYI aaorg “J WE passaquL YG pue Dg pue ‘7 We oss GD pue gy soul] ay ‘seaurffoo axe eam) ou G ‘D‘g *Y Siulod ayy SuoWY ‘9 seeaUyT{O9 ase |. 4, 6, ¥ Siutod ayy “sinav0 (1) azoym asea ayy uy (q) “VD 01 faltered ase §,y pue y,9 10 ‘Dg 0} jayresed ase 9 pue dg 10 ‘gy O} Jajfesed ae A,g pue W,¥ (I) SAjpansadsos “47 pur “,V ¢,,) 18 WaLnduod are (5,y “Y,D‘VD) (OD ‘d, 8 ‘DA “(NAV AV) souy Jo sajden ay (1) :S1NJ9O SUOTIENS SUIMOIJOJ By) JO JuO yseay IY (e) ney) aaosg “Ajaatigadsar ‘g ‘y pue © ‘q swujod souyap suo ‘Kreps *,g ueyy J9yI0 ,5,9,¥ pur ,ggy So[dd19 JO. uONSassaqut Jo 1urod ay? Ay Aq pue Sy URI 18yI0 ,D,g,V pur LV GV Sofauto Jo uorrsasiajut Jo yurod ayy py Aq aiouaqy “Kjaanoodsax ‘gy “VD ‘DA Sopis uo sjutod aq ,D “gf “,V of pue Burin & oq DAV OT “s 2. axSuvjn Jo sof ap pue“rypeay a srpeninoxs sip yo sono 4 pue 7 CQ swawBes Jo s~pSug] a4) aynduroz) “Ajaanoadsar ‘“ajuad0yI0 ayy pur ‘loyuaour ayy ‘rjaounost9 ay aq HY pue ‘7 ‘O 3] DEY [Buen vy -p “snipes ajqissod Jso]jeuls otp Jo ojos e Buope 2 Buneow pue g pue y Sumueyu0o asoyds & yONJSUOD 41 Jo Sapis juaLaZIp UO ¥ pue Y syuTOd om) puL 4 sUEId B DOAID “¢ Anjauouo8uy, pue kyauoen *T rardey a14 Chapter 1. Geometry and Trigonometry ‘This property was proved independently by F. Bolyai (1833) and Gerwin (1835). Its three dimensional version was included by Hilbert in the list of 23 problems that he presented at the International Congress of Mathematicians in 1900. Hilbert stated that this property does not hold for polyhedra and asked for a complete invariant that gives the obstruction to transforming one polyhe- dron into another. The problem was solved by M. Dehn, who constructed the required invariant. le a Figure 1.4.2 Let us prove the Bolyai-Gerwin theorem. First, note that using diagonals, ‘one can cut any polygon into finitely many triangles. A triangle can be trans- formed into a rectangle as in Figure 1.4.2. We showed that two squares can be cut and reassembled into a single square; thus it suffices to show that from a rectangle one can produce a square. Figure 1.4.3 Let ABCD be the rectangle. By eventually cutting the rectangle ABCD into smaller rectangles and performing the construction below for each of them, we can assume that AB/4 < BC < AB/2. Choose the square AXYZ with the same area as the rectangle such that XY intersects CD at its midpoint P (Figure 1.4.3). Let M be the intersection of AB and XY, and N that of AD and YZ. The triangles AZN and AXM are congruent, so the quadrilaterals MBCP and DNY P have the same area. A cut anda flip allows us to transform the second quadrilateral into a trapezoid congruent to the first (the two are congruent, since PC = PD, LDPY = LC PM, and they have the same area). ‘We have proved that any polygon can be transformed into a square. But we can also go backwards from the square to the polygon, and hence we can transform any polygon into any other polygon of the same area, with a square as an intermediate step. 1.4. Dissections of Polygonal Surfaces 15 Show that for n > 6 an equilateral triangle can be dissected into n equilat- eral triangles. An equilateral triangle can be dissected into six, seven, and eight equilateral triangles as shown in Figure 1.44. The conclusion follows from an inductive argument, by noting that if the triangle can be decomposed into n eeqilateral triangles, then it can be decompased into n + 3 triangles by dissecting one of the triangles of the decomposition in four. Figure 1.4.4 1. Given three squares with sides equal to 2, 3, and 6, perform only two cuts and reassemble the resulting 5 pieces into a square whose side is equal to 7 (by a cut we understand a polygonal line that decomposes a polygon into two connected pieces). 2. Prove that every square can be dissected into isosceles trapezoids that are not rectangles. 3. Given the octagon from Figure 1.4.5 one can see how to divide it into four congruent polygons. Can it be divided into five.congruent polygons? Figure 1.4.5 4. Show that any cyclic quadrilateral can be dissected into n cyclic quadri- laterals for n > 4. 5. Show that a square can be dissected into # squares for all n > 6. Prove that this cannot be done for n = 5. 6. Show that a cube can be dissected into n cubes for n > 55.16 Chapter 1. Geometry and Trigonometry 7. Determine all convex polygons that can be decomposed into parallelo- grams. ~ exists a dissection of the polygon into n + 1 convex polygons such that the 2n points lie on the boundaries of these polygons. 9. Let ABC be an acute triangle with ZA = nZB for some Positive integer n. Show that the triangle can be decomposed into isosceles triangles whose equal sides are all equal. 10. Prove that if a certain rectangle can be dissected into equal rectangles similar to it, then the rectangles of the dissection can be rearranged to have all equal sides parallel. : 11. A regular 4n-gon is decomposed into parallelograms. Prove that there exists at least one rectangle in the decomposition. Find the sum of the areas of all rectangles from the decomposition. 12. Find with proof all possible values of the largest angle of a triangle that can be decomposed into five disjoint triangles similar to it. 1.5 Regular Polygons This section discusses two methods for solving problems about regular poly- gons. The first one consists of the use of symmetries of these polygons. We illustrate this with the following fascinating fact about the construction of the regular pentagon. Of course, there exists a classical ruler and compass con- struction, but there is an easier way to do it. Make the simplest knot, the trefoil knot, on a ribbon of paper, then flatten it as in Figure 1.5.1. After cutting off the two ends of the ribbon, one obtains a regular pentagon. . Prove that given any 2n distinct points inside a convex polygon there | 1.5. Regular Polygons ” To convince ourselves that this pentagon is indeed regular, we note that it is obtained by folding the chain of equal isosceles trapezoids from Figure 1.5.2. ‘The property that explains this phenomenon is that the diagonals of the pen- tagon can be transformed one into another by rotating the pentagon, and thus the trapezoids determined by three sides and a diagonal can be obtained one from the other by a rotation. n Figure 1.5.2 The second method employs trigonometry. It refers either to reducing met- ric relations to trigonometric identities or to using complex numbers written in trigonometric form. We exemplify the use of trigonometry with the fotlow- ing problem, which describes a relation holding in the regular polygon with 14 sides. Let AyAzA3... Aja be a regular polygon with 14 sides inscribed in a circle of radius R. Prove that Al A} + Ay A} + Ag A} = 7R?. Let us express the lengths of the three segments in terms of angles and the circumradius R. Since the chords A; A3, A3A7, and Aj Az are inscribed in arcs of measures 1/7, 21/7, and 37/7, respectively (see Figure 1.5.3), their lengths are equal to 2R sin 7/7, 2R sin 2xr/7, and 2R sin 37/7. Hence the identity to be proved is equivalent to 2a 3 4R? (su? ; + sin? 4 sin? >) =7R. Using double-angle formulas we obtain an | 23H 4R? (sw? +sin? 7+ sin? *) 2n 4x =) = 2R?2{1—cos— — cos — + 1 —cos—— =2R"( cos +1 cos + 7 To compute the sum 2n an ern cos + 60s 740 Chapter 1. Geometry and Trigonometry Figure 1.5.3 we multiply it by sin 27/7 and use product-to-sum formulas. We obtain 1 e 47 Sin z 2m gin BH . 4a 1. 20 5 (sin + sin — ~ sin — + sin ~ sin) = 1 gin 2, 27 7 7 7 MS 25 Here we used the fact that sin 8/7 = sin(2x ~ 67/7) = ~ sin 67/7. Hence the above sum is equal to —4, and the identity follows. Here is a list of problems left to the reader. 1. Let ABC and BCD be two equilateral triangles sharing one side. A line passing through D intersects AC at M and AB at N. Prove that the angle between the lines BM and CN is 60°. 2. Let ABCDE be a regular pentagon and M a point in its interior such that LMBA = LMEA = 42°. Prove that LCMD = 60°. 3. On the sides of a hexagon having a center of symmetry construct equi- lateral triangles in the exterior. The vertices of these triangles that are not vertices of the initial hexagon form a new hexagon. Prove that the midpoints of the sides of this hexagon are vertices of a regular hexagon. 4. Let AyA2A3A4A5A6A7 be a regular heptagon. Prove that 1 1 1 th =auxt ; AlAz A\A3” Ai Ag 5. Let AyA2A3AaAsAoA7, Bi BpB3B4BsBeBy, C\C2CyCaCsCoCy be reg- ular heptagons with areas S4, Sp, and Sc, respectively. Suppose that A1A2 = ByB3 = Cis, Prove that 1 Set5c Chen eels 5B ay 1.5. Regular Polygons 6, Let P; PpPs... Pia be a regular dodecagon. Prove that P Ps, P4Ps, and P3P¢ are concurrent. Inside a square ABCD construct the equilateral triangles ABK, BCL, idpoii f the segments KL, LM, M, and DAN. Prove that the midpoints of emi : aa NK and those of AN, AK, BK, BL, CL, CM, DM, and DN are the vertices of a regular dodecagon. x On a circle with diameter AB choose the points C, D, E on one side of AB, and F on the other side, such that AC=C D=BE= 20° and BF= 60°. Let M be the intersection of BD and CE. Prove that FM = FE. @ 9. Let A1AzA3....Azg be a regular polygon with 26 sides, inscribed in a " Gircle of radius R. Denote by A’, A’, and Af the projections of the orthocenter H of the triangle A1A7Ag onto the sides A7Ag, A1Ag, and ‘A,Ap, respectively. Prove that R HA, — HA, +HA,= zr 10. Let AyA2A3... An be a regular polygon inscribed in the circle of center O and radius R. On the half-line OA; choose P such that A; is between O and P. Prove that Tl PA; = PO" — R". f=) . Let AyAzA3... Aang1 be a regular polygon with an odd number of sides, and let A’ be the point diametrically opposed to A2n41. Denote A’Ay = 41, A’A2 = a2, ..., A’An = @n. Prove that Q — a7 4.03 — ag ++ Ey = where R is the circumradius. . Given a regular n-gon and M a point in its interior, let x1, 2, .-., Xn be the distances from M to the sides. Prove that 1 1 1 2n tote ¢oes, ma * x2 + Xn a where a is the side length of the polygon.20 Chapter 1. Geometry and Trigonometry 13. Let n > 2 be an integer. Under the identification of the plane with the set of complex numbers, let f : R? —> R be a function such that for any Tegular n-gon AyA2...An, FAI) + f(A2) +++ + f(A) = 0. Prove that f is the zero function. 14. A number of points are given on a unit circle so that the product of the distances from any point on the circle to the given points does not exceed 2. Prove that the points are the vertices of a regular polygon. 1.6 Geometric Constructions and Transformations We will consider constructions with ruler and compass that lead to an analysis of certain geometric transformations. Here is an example of a construction ‘o be done with a compass with fixed opening. It appeared at the German Mathematical Olympiad (Bundeswettbewerb Mathematik) in 1977. Given three points A, B, and C in the plane and a compass with fixed opening such that the circumradius of the triangle ABC is smaller than the opening of the compass, construct a fourth point D such that ABCD is a Parallelogram. The problem requires us to translate the point A by the vector BC. Let a be the length of the opening of the compass. The construction is very simple in the particular case where AB = BC = a. Indeed, if we constfict the two circles of radius a centered at A and C, one of their intersections is B and the other is the desired point D. Our intention is to reduce the general case to this Particular one. We have already solved the problem when both the segment and the vector have length equal to a. Let us show how to translate A by the arbitrary vector BC when AB = a. The restriction on the size of the triangle implies that there is a point P at distance a from B and C. Construct P as one of the intersections of the circles centered at B and C with radii a. The point D is obtained from A by a translation of vector BP followed by a translation of vector PC, both vectors having length a (Figure 1.6.1). If AB has arbitrary length, construct Q such that QA = QB = a. Then eae @ to R by the vector BC and finally translate A to D by the vector OR. Here are more problems of this kind. ic 21 1.6. Geometric Constructions and Transformations A — P Figure 1.6.1 1. Given a polygon in the plane and M a point in its interior, construct two " points A and B on the sides of the polygon such that M is the midpoin of the segment AB. 2. Given a polygon in the plane, construct three points A, B, C on the sides ~ of this polygon such that the triangle A BC is equilateral. 3. With ruler and compass construct a trapezoid given the lengths of its sides. 4. Construct a trapezoid given the lengths of its diagonals, the length ot the . line segment connecting the midpoints of the nonparallel sides, and one of the angles adjacent to the base. 5. Let C; and C2 be two concentric circles. Construct a line d that intersects these circles consecutively in A, B, C, and D, such that AB = = cD. 6. The lines m and n pass through a point M situated equidistant frome : parallel lines. Construct a line d that intersects the four lines at A, B, C, D such that AB = BC =CD. 7. Given two arbitrary points A and B, with a compass with eee construct a third point C such that the triangle ABC is equilateral. 8. Given two arbitrary points A and B, with @ compass with fixed opening construct a point C such that ABC is a right triangle. 9. Given a circle in the plane, construct its center using only a compass.22 Chapter 1. Geometry and Trigonometry 1.7 Problems with Physical Flavor 1.7. Problems with Physical Flavor In this section we discuss some elementary problems that can be easily solved at the physical level of rigor, and we explain how physical intuition helps us find a mathematical solution. We exemplify this with the problem of the Tori- celli point, and leave the rest of the problems as exercises. Find the point in the plane of an acute triangle that has the smallest sum of the distances to the vertices of the triangle. The idea of Leibniz was to place the triangle on a table, drill holes at each ! vertex, and suspend through each a ball of weight equal to 1 hanging on a thread, then tie the three threads together (Fig 1.7.1). The system reaches its : Figure 1.7.1 Lex. D yFigure 1.7.2 iron consider a vector of length nu- ‘ lyhedi 2. Orthogonal to each face of a poly! oc tgcutvads, Prove that me merically equal to the area of that face, sum of these vectors is zero. i ihedral angles of a tetrahedron that the sum of the cosines of the Gihedr s a ove at exceed 2; moreover, it equals 2 if and only if the faces of the equilibrium when the gravitational potential is minimal, hence when the sum of tetrahedron have the same area. the lengths of the parts of the threads that are on the table is minimal. The point P where the three threads are tied together is the one we are looking for. On the other hand, the three equal forces that act at P, representing the weights of the balls, add up to zero, because there is equilibrium; hence APB = ZBPC = ZAPC = 120°. This way physical intuition helped us locate the point P. Let us row prove rigorously that if APB = LBPC = APC = 120°, then AP + BP +CP is minimal. Let D be such that BCD is equilateral and such that BC separates A and D (Fig 1.7.2). By Pompeiu's theorem (Section 1.1), we have BQ +CQ > QD for any point Q in the plane, with equality if and only if Q is on the circumcircle C of the triangle BCD. By the triangle inequality, AQ + QD > AD, with equality if and only if Q is on the line segment AD. Hence AQ + BQ +CQ > AD, and the equality is attained only if Q coincides with the intersection of C and AD, that is, with point P. It - follows that P is a minimum and the only minimum for the sum of the distances to the vertices, and we are done. We invite the reader to consider the three dimensional version of the Toricelli point (problem 4), along with some other problems of the same type. assume there is a point P in its interior om P to the vertices is minimal. ZAPB and LCPD are supported al to the line determined 4. Let ABCD be a tetrahedron and such that the sum of the distances fr Prove that the bisectors of the angles by the same line. Moreover, this line is orthogon: by the bisectors of LBPC and LAPD. lyhedron, show that there exists a face . ea 1 5 Gin hic such that ep jection of the point onto the plane of of the polyhedron such that the pr that face lies inside the face. aight river. In what place should the length of the road AMNB? the river, and the 6, The towns A and B are separated by a str onstruct the bridge MN to minimize (ite bridge is supposed to be orthogonal to the shore of river is supposed to have nonnegligible width). 7. Five points are given on a circle. A Perpendicular s drawn fhrovgh the ‘id of the triangle formed by any three o! t a ° ing the remaining two. Such a perpen cular is drawn for each triplet of points. Prove that the ten lines obtained in this way have a common point. Generalize this statement ton points. 1. Consider on the sides of a polygon orthogonal vectors of lengths propor- tional to the lengths of the sides, pointing outwards. Show that the sum of these vectors is zerd.24 Chapter 1. Geometry and Trigonomenry; 8. Find all finite sets S of at least three points in the plane such that for all distinct points A, B in S, the Perpendicular bisector of AB is an axis of symmetry for S. 9. The vertices of the n-dimensional cube are m-tuples (x1,.x2,.-. , Xn) where each x; is either 0 or 1. Two vertices are joined by an edge if they differ by exactly one coordinate. Suppose an electrical network is Constructed from the n-dimensional cube by making each edge a 1 ohm resistance. Show that the resistance between (0,0,...,0) and (41,2. 1)is Bo R= >> my" jal 18 Tetrahedra Inscribed in Parallelepipeds There is no equilateral triangle in the plane with vertices of integer coordinates. In space however, the tetrahedron with vertices (1. 0,0), (0, 1, 0), (0. 0. 1), and G, 1, 1) is regular. Thus there is a regular tetrahedron inscribed in the unit cube as shown in Figure 1.8.1. Figure 1.8.1 Asa consequence there is a short way of computing the volume of a regular tetrahedron whose edge is equal to @. All we have to do is subtract from the volume of the cube with edge equal 10 a/2/2 the volumes of the four tetra- hedra we cut out. So the volume of the regular tetrahedron is a34/2/4 — 4 x 6/2/24 = BSI/12. By deforming Figure 1.8.1 we see that any tetrahedron can be inscribed in a parallelepiped in this way. We are interested in the relationship between the Properties of the tetrahedron and those of the associated parallelepiped, Two eT i 25 1.8, Tetrahedra Inscribed in Parallelepipeds will be considered, that of the orthogonal tetrahedron and that of the cases. isoscel hedron. . | ; ‘eerahedron is called orthogonal if the opposite edges are orthogons)- aa ently, a tetrahedron is orthogonal if and only if the ssociated ea ra Coeeed is rhomboidal, ice.,, all of its faces are rhombi. Ineo 1c a ee sonal if and only if the two diagonals of each face ne pal eniped wee thogonal. On the other hand, a parallelogram whose diag are or |. i i bus, hence the conclusion. ; tree ele tetrahedron is one for which the opposite edges are ome * vlines called equifacial, since its faces are congruent triangles. ee hed is isosceles if and only if the diagonals of each face of the asso ated seealclepiped are equal, hence if and only if the associated parallelepip par > 3 ne t us show how these consitlerations can be applied to solve the polowing ier given in 1984 at a Romanian selection test for the Balkan Mat problem, n ical Olympiad. : » d2, and d3 be the common perpen- ic va ATaB va (eaten De AD and BC, respectively. Prove that Se hotion is isosceles ifand only if dy, da, and d3 are pairwise orthogo- nal. - ; For the solution, note that if we inscribe the tetrahedron in 3 paralclepiped as in Figure 1.8.2, then dy, dz, and dy are orthogonal to the agonal onthe three pairs of faces, respectively, and hence they are ortl gonad the fares. The parallelepiped is right if and only if every two faces sh ing 3 coe ay edge are orthogonal. On the other hand, two planes are orthogonal fan only if the perpendiculars to the planes are orthogonal. Here paral right if and only if dh, dz and dy are pairwise orthogonal, fro follows. 4, Figure 1.8.2The problems below describe properties of orthogonal and isosceles t hedra that can be proved using the same technique. 1, Express the volume of an isosceles tetrahedron in terms of its edges. 2. Express the circumradius of the isosceles tetrahedron in terms of its edges, 3. Let ABCD be atetrahedron and M, N, P, Q, R, and S the midpoints AB, CD, AC, BD, AD, and BC, respectively. Prove that the segments. MN, PQ, and RS intersect. 4. Prove that if two Pars of opposite edges in a tetrahedron are orthogonal, then the tetrahedron is orthogonal. 5. Prove that ina rhomboidal parallelepiped A,B, C1 Dj AzB)C2 D> the com- mon perpendiculars of the pairs of lines AyC; and By,D2, A, By and 1D2, Ay D2 and BoC, intersect. 6. Prove that in an orthogonal tetrahedron the altitudes intersect, Let ABCD be an orthogonal tetrahedron, Prove that AB? + CD? = AC? + BD? = AD? + Bc? 8. Express the volume of an orthogonal tetrahedron in terms of its edges. ns . Prove that if all four faces of a tetrahedron have the same area, then the tetrahedron is isosceles. ~ 10. In a tetrahedron all altitudes are equal and one vertex projects orthogo- nally in the orthocenter of the opposite face. Prove that the tetrahedron is regular. 1.9 Telescopic Sums and Products in Trigonometry DFe+)— Fe] kel at 1.9. Telescopic Sums and Products in Trigonometry cancel out terms to obtain F(a + 1) — F(A). Similarly for products, and then ree — Far) ~ FO A F@ Fl) . n kx. “, One example is the computation of Dj-1 coskx. Assuming that x 4 2mm, integer, ‘ve multiply by 2sinx/2. From the product-to-sum formula for the an integer, product of a sine and a cosine we obtain (on +2) 8a (-2)*) i s)x—sinfk-s)x 23 in 5 caskx = ¥0(sin(k+ 5) sin( 3 k=l 2 kd 1 a! =sin(n 45) x—singe i 4 , whea x = ‘Thus the initial sum is (sin@ + $)x)/(2sin(x/2)) — 7. Of course, wl 2mxc, man integer, the answer is n. function. For the second example we use a formula for the tangent Evaluate the sun 1 —1 fan eka’ 2s R+k+ where tan—! stands for the arctangent function. 1 In the solution we will use the subtraction formula for the tanget tana —tanb tana — 8) = Tana tan” which gives the formula for the arctangent a yp u~v tan! uy — tan“! y = tan“! -1 For simplicity, set ay = tan! k, Then at EE = santas — a4). Rake) 14+kk4+1) Hence the sum we evaluate is equal to Sean! (tan(ax41 — a2) = dew — %) = fost —ay m stan (n+ 1). Similar ideas can be used to solve the following problems.28 Chapter 1. Geometry and Trigo 1. Prove that sinx | sin2r sinnx cosx * cost b+ Gogg weoex — SEDs sinx cos’ x" for all x £kZ, kan integer. 2. Prove 1 1 1 cds 1° see ‘cos 0° cos 1° io Hee ee cos 1° cos 2° Cos 88° cos 89° sin? Jo" 3. Let n be a positive integer and a a real number, such that a/z jis j b is an irrational number. Compute the sum 1 cosa —cos3a * cosa—cossa t 7 1 +-——___ coSa — cos(2n + 1)a° 4. Prove the identity 2 1 Dotan = tant" isl 2K? ntl 5. Evaluate the sum Y jun’ . mie 2 where a £kn,k an integer. 6, Prove that co 3" sin? $ 1 For = 7 (a -sin a). 7. Prove that the aver . cotl*, ‘age of the numbers nsinn®, n = 2,4,6,... , 180, is 8. Prove that for ever itive ii TY positive integel 5 (m =0,1,...,n,kan integer), 11 and for every reat number x # 4 I 1 => ++. 1 _ sindx + indy + 7°" + Gaga = Cotx — cot 2x, 110. ‘Trigonometric Substitutions 9, Compute tan1 + tan2 eet tan 2” cos2 cos4 cas 2t+t” 10. Prove that for any nonzero real number x, cox _ sinx 1). Prove that for any integer 7 > 1, 2x cos 4n xn 1 1 2 ~ cos 12. Evaluate the product 43, Let n be a positive integer, and let x be a real number different from PCL 4 It), k= 1, 2,..- +n, Lan integer. Evaluate the product T1(-20085)- 14, Prove that ft 2x -3* TC 4200532) =1 1.10 Trigonometric Substitutions Because of the large number of trigonometric identities, the choice ofa clever trigonometric substitution often leads to a very simple solution. This is the case with all the problems presented below. The substitution is usually suggested by the form of an algebraic expression, as in the case of the following problem. Find all real solutions to the system of equations xei3r=y, y—3y=z, -3z=x.~ Chapter 1. Geometry and Tri 2cos3u = 2cos v, 2cos3v = 2cosw, 2cos3w = 2cosu, By using of the triple-angle formula for both cos 3u and COs v, the first equation becomes cos 9u = cos 3v. Combining this with the second equation, we obtain cos 9u = cosw. As before, cos27u = cos 3w, and the third equation yields cos27u = cosu. This equality holds if and only if 274 = 2kr -. u for some’ integer k. The solutions in the interval [0, J are u = kr/i4,k =0,1,...,14 and u = kw/13,k =1,2,... 12. Consequently, x=2coskn/14, y= 2cos3kr/14, z= 2cos9kx/14, k=0,1,...,14, * and x = 2coskr/13, y = 2cos3kn/13, z= 2cos 9k /13), k= 1,2,...,12. are solutions to the given system of equations, Since there are at most 3 x3x 3 = 27 solutions (note the degree of the system), and we have already found 27 distinct ones, these are all the desired solutions. We Proceed with an example where the tangent function is used. Let {Xp}, be a sequence. ‘Satisfying the reccurrence 1 V3x_—1 >t Xn +3” , Prove that the sequence is periodic. Recall the formula for the tangent of a difference: tana ~ tanb tan(a — b) = T—tanawanF" Note also that tan $ = a If we rewrite the recurrence relation as 31 1.10. ‘Trigonometric ‘Substitutions tant, for some real number t. Then x2 = icf 2 a aincly 2 6) > 1. Since the ne inductively Xq = tan(t — (1 — 1)7/6), n= en at 2c of period mr, we obtain Xn = Xn46, Which shows that the sequence has period 6. For what values of the real parameter @ does there exist a real number x 1 satisfying Vi-x2?2za-x? 2. Given four distinct numbers in the interval (0, 1), show that there exist . Gi two of them, x and y, such thai 1 2 =. o ind all 3. The sequence [Xn}a Satisfies /%n42 4 2 < xn < 2 foralln zi Find al possible values of x19s6- 4, Find all real solutions to the system of equations 2+xy=y, 2ty2=2, Qe+2x =x. 5. Find all real solutions to the system of equations 6. Prove that (x+y) — xy) 1 +2?) +y?) forallx,y €R. 7. For each real number x, define the sequence {x,},, recursively by x1 = x, “ 1 1 wml Ean . ii (for X41 would be = :H1, then the sequence terminates (1 nd cay How many such sequences terminate after the eighth term?“nape nero ote 32 Chapter 1. Geometry and Trige 8. A sequence of real numbers a}, a2, 43,... has the property that a4 (kag + 1)/(K — ay) for every positive integer k. Prove that sequence contains infinitely many positive terms and infinitely m: 9. Given -1 i=1 10. Let x9 = O and x4,.x2,... 5 > O, with P%_; xg = 1. Prove that a xk T ae FAV F0+ Px Ry 2 11. Find all triples of numbers x, y, z € (0, 1), satisfying tyes? 4 dry = 1. 12. Let a, b, and c be given positive numbers. Determine all Positive numbers x, y and z such that X+y+z=a+bt+e, xyz — (ax + By +.c7z) = abe. Chapter 2 Algebra and Analysis34 Chapter 2, Algebra and Analysis 2.1 No Square is Negative In this section we will consider some applications of the simplest inequality in algebra: x? >0, where equality holds if and only if the real number x = 0. Herc is the first example. Let a be a real number. Prove that 4a — at < 3. This problem was proposed to young students, who did not even skim through a calculus book. The inequality is equivalent to @ —1)?+2(a-1)? > 0, which is clearly true. The second example is from the Romanian Mathematical Olympiad 1981, proposed by T. Andreescu. Determine whether there exists a one-to-one. function f ; R — R with the property that for all x, 1 F(x’) — Pe@y> z ‘We will show that such functions do not exist. The idea is extremely simple. We look at the two numbers that are equal to their squares, namely x = 0 and x = 1, for which f(x) and f(x?) are equal. Plugging x = 0 into the relation we obtain FO ~ PO = ; Moving everything to the right side yields o> (10-3). This implies f(0) = 3. Similarly, ff we plug in x = 1, we obtain f(J) = } which is the same as f (0), and thus F cannot be one-to-one. We list below a number of problems that can be solved similarly, 1. Leta, b,c, and d be reat numbers. Prove that the numbers a— 22, b—c?, ¢—d?, and d — a? cannot all be larger than 2. 2.1. No Square is Negative 2. Find all real solutions to the system of equations xty=Vvaz—1, ytz= Vax -1, ztxaV4y—1. 3, Let x,y be numbers in the iterval (0, 1) with the property that there ” Gxists a positive number a different from 1 such that log, a + logy @ = 4log,, a. Prove that x = y. . 4, Find all real triples (x, y, z) that satisfy x4 + y4 + 24 —4xyz = —1. 5, Determine all triples of real numbérs x, y, z satisfying Qry-2 21, z-le+yl2-1. 6. Let a, b, and c be real numbers such that a? +c? < 4b. Prove that for all x ER, x4 + 4x3 + bx? +ex4+1>0. 7. Prove that for all real numbers x, y, z, the following inequality holds: 3 2 4 4? — ay ye 3222 GR 8. Find the real numbers x1, x2, ... Xn Satisfying 1 Vey = 12 42x. — Bo eng 0? = 3 $32 +--+ Hn) 9. (a) Let a, b, ¢ be nonnegative real numbers. Prove that ab + be +ca > J3abc(a +b +0). {b) Let a, &, c be nonnegative real numbers such that a + b+e= Prove that a4 4c 4+ Si2abe <1. 10. Determine f : N -> R such that for all k, m. 7, Fk) + Fen) — FU) Flan) > 1.36 Chapter 2, Algebra and Analysis | 11. Let a, b, c be the lengths of the edges of a right parallelepiped and let d be the length of its diagonal. Prove that 0? + Be? 4 a > abcd J3. 12. Prove that ina triangle ABC the following inequality holds: cos A cos BeosC < i (3. Ifat, ag, ... , ay are real numbers, show that Lv eosta aj) > 0. i=l jot 2.2 Look at the Endpoints This section contains inequalities that could be proved by using the Property that certain real functions reach their extrema at the endpoints of the interval of definition. Two kinds of functions are considered: linear functions, which have both extrema at the endpoints of their domain, and convex functions, whose maximum is attained on the boundary of the domain, The idea is to view an expression as a linear or convex function in each of the variables separately, and usc this to bound the expression from above or below. ‘We exemplify these ideas with a Problem that appeared at a Romanian selection test for the International Mathematical Olympiad in 1980. For a given positive number a find the maximan of n Do —a1)(a— a2) ++. — ay seul — ayss)---(a— ), kat where a1, an, ..., , range independently over the interval {0, a]. For an index k, fix aj,a,... 2 Ok-1s ket, ~-+ Gn, and think of the given expression as a function in ag. This function is linear, hence its maximum on the interval [0, a] is attained at one of the endpoints. Repeating the argument for each variable, we conclude that the expression reaches its maximum for a certain choice of a = 0 or a, where k = 1,2,... a. If all ay’s are equal to 0, or if two or more a;’s are equal to a, the sum is 0. If one ay is @ and the others are O, the expression is equal to a”; hence this is the desired maximum. Here is an example that appeared at the W.L, Putnam Mathematical Com- Petition and is solved using convex functions. Endpon 37 2.2. Look at the ints Let n be a natural number and let x; © (0,1), i = 1,2,--. +7 Find the i 2and 0 0, prove that - 2 Grbrerdra(septiey +s) ) (Za) = 2.3 Telescopic Sums and Products in Algebra We have already seen in Section 1.9 that Many sums can be easily computed by telescoping them, that is, by putting them in the form a : DFG +1) — FL k=l Indeed, in such a sum the F(k)’s, for k between 2 and n — 1, cancel out, leaving the sum equal to F(n + 1) — F(1). The reader might notice the similarity between this and the fundamental theorem of calculus, and conclude that we are finding a discrete antiderivative for the terms of the sum, A simple example employing the method of telescopic summation is the computation of Yp_ k!-k. If we write k!-k = k!-(k-+1—1) = e-4+1)!—Ht, then the sum becomes Deel + D!- ky, which, after cancellations, is equal to(n+1)!-1. Evaluate the sum n yx I bal KA DE ARSED ‘The idea is to rationalize the denominator. We have (+ DV = AEFI + VE + RET) =kKk+ 1-4 DP HELD. . * 39 23, Telescopic Sums and Products in Algebra 3 ‘The sum becomes 1 n 1 1 =1l— . Sekt DET VEEL (4-gen) =! Ti k=l kk+) A similar method can be used for computing products. In this case one writes the expression as a product of fractions whose numerator n denom me tors cancel out alternately to leave the numerator e ction vjenominator of the last, or vice versa. Here is an example. Prove that Fi(-3)-4 4 a=? Truncating the product and writing each factor as a fraction, we obtain fi(-4)-f(-) 0d) fe 1 N+1_N+1 =N 2 aN Letting N tend to infinity we obtain that the product is equal to } We invite the reader to solve the following problems. 1. Compute 77_, KIG? +4 +1). 2. Let a1, a2,... ,@n bean arithmetic progression with common difference d. Compute Z 1 Kai ARE 3. Evaluate the sum » s Gena GEA — Qe’ 4 GF- 2G) att) 4. The sequence {%q}n is defined by xy = 4 and xe41 = x} +m. Find the Breatest integer less than 1 1 1 — ++. nti’ x41 100 +1‘0 , Chapter 2, Algebra and Anal, a, On an Algebraic Identity 41 5. Let F, be the Fibonacci sequence (F, = =L1FR=1, Fur =iyt Fy Evaluate 42. Prove that wad wl wi ime co ® yy Fn Fai ; n=2 1 © LU RGEn . n=2 13. Compute the product Hi( veh). n=0 2 6. Compute the sum 2, Lz = 1, and Lny2 = Lngi + Ln, for n > 1, be the Lucas 14, Let Li = ltptgstyitgeae.. +) got sequence. Prove that m 7. Prove the inequality [ [fon = Fens ket 1 1 1 sete te Vi+43 "Bait * Tear. ames 8. Prove the double “say Wane 1 Vii) < Fat yet ty te <2 Va=D, > 24, where {Fy}, is the Fibonacci sequence. 2.4 Onan Algebraic Identity This section is about problems that can be solved using the following algebraic identity: 1 1 9. Let Xe = OP bart SYD AY 4 5Y). . k % = ——— se K- YB ERB LED Here are two examples. Evaluate the sum Prove that a1 + a2 +--+ +4 ag99 < 50. 10. Prove the inequality y 4k Byaeet By using the above identity we obtain "(22 4 2k +1) — (2k? — 2k +1) co 1 Larne (——__ - —_ yt =) (gear mac) na Fn — 1 Tal where F,, is the mth term of the Fibonacci sequence. and we are done.42 Chapter 2, Algebra and Analy; Given ann x n matrix A with the property that A? = 0, prove that t matrix M = 4A? + A+ In is invertible. { Indeed, the inverse of this matrix is $A? — A + Ip, since the product the two gives }A* + [,,, which is equal to the identity. Of course. in this c. the conclusion can also be detived from the fact that the eigenvalues of M ar; nonzero, which is a direct consequence of the spectral mapping theorem. We give below more examples that make use of this identity. . Prove that for every integer n > 2, the number 22"~? + 1 is not a pri number. 2. Prove that any sequence satisfying the recurrence Angie Xn = V Oi is periodic. : . 3. Compute the sum 4. Evaluate 5. Show that for any 7, there exist infinitely many numbers a such that the number n4 + a is not prime. 6. Show that 14+ 4” is a prime number if and only if 2 == 1. 7. Consider the polynomial P(X) = X4 + 6X? — 4X + 1. Prove that P(X‘) can be written as the product of two polynomials with integer coefficients, each of degree greater than 1. 8. Prove that for any integer n greater than 1, n!? +64 can be written as the product of four distinct integers greater than 1. 9. Let m and n be positive integers. Prove that if m is even, then m Learner 0 is not a prime number. 25. Systems of Equations 43 2.5 Systems of Equations In this section we have selected non-standard systems of equations. The first problem is from the 1996 British Mathematical Olympiad. Find all solutions in positive real numbers a,b, c,d to the following sys- fem: atb+ctd=12, . abed = 21 + ab tae + ad + be +bd + ed. Using the AM-GM inequality in the second equation, we obtain abed > 27 + 6Vabed. Moving everything to the left and factoring the expression, viewed as quadratic polynomial in fabed, yields (Vabed + 3)(Sabed — 9) 2 0. This implies Vabed > 9, which combined with the first equation of the system gives b d Yabed > ates The AM-GM inequality implies that a = b = ¢ = d = 3 is the only solution. ‘The second problem is from the 1996 Vietnamese Mathematical Olympiad. Find all positive real numbers x and y satisfying the system of equations 1 14——)=2, va( + =) 1 1-——} =4v2. V9(1-35) It is natural to make the substitution ./x = u, fy = v. The system becomes45 Chapter 2. Algeb: Systems of Equations apt Igebra and Anal| 25. absolute value of the complex nu 3. Find all solutions to the system of equations But u? + v? is the square of the on he . . Fi u + iv. This suggests that we add the second equation multiplied by i to at first one. We obtain aey1~™ 2 4 Hoe V2 Ye wtivt oe uf). 4y? +1 42 The quotient (uw — év)/(u? + v2) is equal to Z/lzl? = E/(e2) = 1/z, so ai above equation becomes 4, Find ax) + by? if the numbers a, b, x and y satisfy the system of equa- tions Hence z satisfies the quaiiratic equation ax +by = 3, 2:{2 , 472 ax’ + by =7, . [ss t+i— }z+1=0 3 Vi Vi ax? + by = 16, with solutions ax + by! = 42. 1 2 , [2v2 (Z * a) " (35 * “) , 5. Solve the system where the signs + and ~ correspond. x42 =2y, This shows that the initial ‘system has the solutions x 2 2 = a2 1 2 yt . = (52H): ye 22 : > V3 V2, v7 24 Z+—=2x. z where thé signs + and — correspond, The systems below are to be solved in real numbers, unless specifies otherwise. 6. Solve the system of equations 1. Solve the system of equations @+yPen xt loge t+ V2 +1) = y, OF =x, @txP=y. y+ logy t fy? +1) =z, [41 = 7 lou + V2? +1) = 2. 7. Solve the system 2. Solve the system be = Iyzl, y — Il = lel, 2 kl = byl Jog(2xy) = logx logy, log(yz) = log y logz, log(2zx) = logz logx.46 Chapter 2, Algebra and Analysis 8. Find the solutions to the system of equations x+y) + f= 1.1, be + 6}4+2=22, G}+y+ lz] = 3.3, where | | and {} denote respectively the greatest integer function and the fractional part function respectively. 9. Fora given complex number a, find the complex solutions to the system (Q + 92 +.3)x4 =a, Gi + x2 + x4)x3 =a, \ Orr +23 +24)x2 = a, : G2 +33 4x4)a1 =a. 10. Find all real numbers a for which there exist nonnegative real numbers X1, X2, X3, X4, X5 Satisfying the system 3 Vix =a, iat 5 Vita = 0, fet 5 bx =a. ket IL. Solve the system of equations x3 —9(y? + 3y—3) =0, y —9@? + 32-3) =0, 2-90? 43x—-3) =0, 12. Solve the system : ax thy = (xy, by +ez = (y- 2), ez+ax=(z—x)’, where a, b,c > 0. 26. Periodicity 47 13. Let a, b, c be positive real numbers, not all equal. Find all solutions to the system of equations in real numbers x, y, 2. 2.6 Periodicity Periodicity plays an important role in mathematics, and for this reason we have included some problems involving it. The first example shows how periodicity can be used to give a short proof of the Hermite identity for the greatest integer function: 1 tit |xe tent ae =u. for all x € Randn €Z. ‘The proof proceeds as follows. Define f : R-> N, s * |= vex n pops tale |e dt] eet [ae One can easily check that f(x) = 0 for x € [0, 1). Also, p(x+t)= [rat feet e 2S |+ [ark] -tex ent = soo. Hence f is periodic, with period 4. ‘This shows that f is identically equal to 0, which proves the identity. The second example concems the computation of a sum of binomial coef- ficients. Compute the sum (Cr) +C2)- C3) +48 Chapter 2, Algebra and Analysi Denote our sum by S,. Then _ _{*\_[m-1), (n-2 -3 SGI) _f(a-t -2)_fn~3\ [n~4 (") ae )- (3 )(3 )o~ mn ~2 —3 4) fn- + _ - Le ("0’) (Cr) (3)+ : Since (¢) = ("G1), the first terms of S, and S,4 cancel. If we group the Ath term of S$, the Ath term of S,_. and the (k — 1)st term of S;_2, we obtain cp (G) ~ (2) — C21), which is zero by the recurrence formate, for; binomial coefficients. Hence all terms cancel. it follows that S; — Sy 1 + Sn-2 = 0. Add Sy_1 ~ S,_2 + Sy..3 = O to this equality and get S, = Spade fi This shows that 5, is periodic of period 6. j So it suffices to compute S,, for n = 1, 2, 3, 4, 5, and 6, the other values * repeating with period 6. We obtain Sénu1 = St = 1. Sonar = > = 0, Son43 = Ss = —1, Sona = Si = =I, Sinus = 0, and Sen = 55 = 1. Here are more problems involving periodicity, 1. Let f : R > R ~ (3} be a function with the that ists | on Ouhne property that there exists f&)—5 F@yay frallxeR. f@toa= Prove that f is periodic. 2. Let f :R + R be a function that satisfies @ F@ +y)+ F&—y) =2F6)F(), forall x,yER; (b) there exists xq with f (x9) = —-1. Prove that f is periodic. 3. Ifa emtion J = R -> Ris not one-to-one and there exists a function g: Rx R-> Rsuch that f(x +y) = g(/(x), y) forall that fis porone 8(/@), ») forall x, y € R, show 4. Let f : R — R bea function satisfying 1 fe +a)=54 YF) ~ soy? for all x € R and acertain fixed a, (a) Prove that f is p iodic, (b) In the case a = 1 give an example of such a function. 26. Periodicity 49 5. The sequence {aq}n is given by 0 < ag < ay + a) < Land Prove that there are real numbers b such that Jan| < b for all n. Find alla € R with the, property that a sequence satisfying the recurrence Anti + Xn—1 = OX, 1 EN, is periodic. Let f : R — Rbea periodic function such that the set {f(m) | n € N) has infinitely many elements. Prove that the period of f is irrational. 8. Let by be the last digit of the number P4243 4.40" Prove that the sequence (b,), is periodic with period 100. Let a, b be positive integers with a odd. Define the sequence ftn}n as follows: no = b, and forn € N, be tay =] Blin if ttn is even mel Uy, -+a_ otherwise. (a) Show that u, R bea monotonic function for which there exist a, b, c,d € R with a # 0, c £0 such that for all x the following equalities hold: x4J5 x+VE f fWOdt=ax+b and f{ f@Mdt =ex+d. i Fi Prove that f is a linear function, i.c., f(x) = mx-tn for some m,n € R.50 Chapter 2. Algebra and A 2.7 The Abel Summation Formula Let ay, a2,... a, and by, b, » bn be two finite sequences of mi Then aby + agbe+++-+ agby = (Qi — 42)b; + (a2 — a3) (by +b) bee Heros — n)br Bat Unt) + ab + Ba ie + by). The similarity with integration by parts is obvious if one temembers thes integration corresponds to the summation of all terms, while differentiati corresponds to the subtraction of consecutive terms. The next problem is a. direct application of this formula. Consider a polygonal line PyP,....P, such that LPyP;P2 = LP, PpPy = 4Pn—2Pn—1Pr, all measured clockwise. If PyPl > P\P) » <->? Pr—1 Pn; show that Py and Py cannot coincide. For the solution, Jet us consider complex coordinates with the origin at Po and the x-axis the line PyP;. Let a be the angle between any two consecutive} segments, and let a) > az > --. > @, be the lengths of the segments, If we set z = e'), then the coordinate of Py is ay +a9z 4 --. 4 ayz"—". We must Prove that this number is not equal to zero. Using the Abel summation formula we obtain Ay bane +o ay2" = (ay ~ a) + (a = 31 4.2) 4-0 Fan tz4--- $274). If. is zero, then this quantity is a strictly positive real number, and we are done. If not, multiply by 1 — z to get (a; — a2)(1 — 2) + (a2 —a3)(1 — 22) +++-F4n(1 —2"). This expression cannot be zero. Indeed, since Jz] = 1, by the triangle inequality we have Maa — a2)z + (az — a3)2? +--- 4a_2"| < Mas — a2)2] + 2 — a3)2"] +--+ + Jag" = (a — a2) + (G2 — 03) + --- +n, and the concJusion follows. * a 27. The. Abel Summation Formula . The following problems can also be solved by using the above summation. formula. 1, Using the Abel summation formula, compute the sums “@I+ 2g t 3g be tng), (b) 14494992 +---+07Qht. ; 4 id by > by = + = in > mmbers a) > G2 2 +-*-Z Gn > Oant 2 TUisty oy > Dys yaa > by tba, --- 21-02 + tan 2 bythe + + + Bq. Prove that for every positive integer k, ah tak +-- tak > bE + + +B 3. Leta, b, ¢, and d be nonnegative numbers such that a S latb<5, atbte yis X12 2 YiYoeeo-0 MaDe An ZV Tn Prove that xy +22 -+++++%n 2 M+ Yate + In fn iti imbers. tforalln, ) a 2 6. Let {a,}, be a sequence of positive nu such that for y a/n. Prove that for all n, eo) Vig bade (145+ a) i=l 7. Let @ : N—> Nbe an injective function. Prove that for all eN HO) Ol Le be 8. Leta; +02 +43 +44-+--- bea convergent series. Prove that the series at + 42/2 +43/3 + 44/4 +--+ is also convergent.