Baselproblem
FromWikipedia,thefreeencyclopedia
TheBaselproblemisaprobleminmathematicalanalysiswithrelevancetonumbertheory,firstposedbyPietroMengoliin1644andsolvedbyLeonhardEulerin
1734[1]andreadon5December1735inTheSaintPetersburgAcademyofSciences.[2]Sincetheproblemhadwithstoodtheattacksoftheleadingmathematicians
oftheday,Euler'[Link],andhisideasweretakenupyears
laterbyBernhardRiemanninhisseminal1859paper"OntheNumberofPrimesLessThanaGivenMagnitude",inwhichhedefinedhiszetafunctionandproved
[Link],hometownofEuleraswellasoftheBernoullifamilywhounsuccessfullyattackedtheproblem.
TheBaselproblemasksfortheprecisesummationofthereciprocalsofthesquaresofthenaturalnumbers,[Link]:
Thesumoftheseriesisapproximatelyequalto1.644934 [Link](inclosedform),aswellasaproof
2
[Link] [Link]
6
justifiedatthetime,althoughhewaslaterprovencorrect,anditwasnotuntil1741thathewasabletoproduceatrulyrigorousproof.
Contents
1 Euler'sapproach
2 TheRiemannzetafunction
3 ArigorousproofusingFourierseries
4 Arigorouselementaryproof
4.1 Historyofthisproof
4.2 Theproof
5 Seealso
6 References
7 Notes
8 Externallinks
Euler'sapproach
2
Euler'soriginalderivationofthevalue 6 essentiallyextendedobservationsaboutfinitepolynomialsandassumedthatthesesamepropertiesholdtrueforinfinite
series.
Ofcourse,Euler'soriginalreasoningrequiresjustification(100yearslater,KarlWeierstrassprovedthatEuler'srepresentationofthesinefunctionasaninfinite
productisvalid,bytheWeierstrassfactorizationtheorem),butevenwithoutjustification,bysimplyobtainingthecorrectvalue,hewasabletoverifyit
[Link].
TofollowEuler'sargument,recalltheTaylorseriesexpansionofthesinefunction
Dividingthroughbyx,wehave
UsingtheWeierstrassfactorizationtheorem,itcanalsobeshownthatthelefthandsideistheproductoflinearfactorsgivenbyitsroots,justaswedoforfinite
polynomials(whichEulerassumed,butisnotalwaystrue):
sinx
Ifweformallymultiplyoutthisproductandcollectallthex2terms(weareallowedtodosobecauseofNewton'sidentities),weseethatthex2coefficientof x
is
sinx 1 1
Butfromtheoriginalinfiniteseriesexpansionof x ,thecoefficientofx2is 3! = 6 .Thesetwocoefficientsmustbeequalthus,
Multiplyingthroughbothsidesofthisequationby2givesthesumofthereciprocalsofthepositivesquareintegers.
TheRiemannzetafunction
TheRiemannzetafunction(s)isoneofthemostimportantfunctionsinmathematics,[Link]
functionisdefinedforanycomplexnumberswithrealpartgreaterthan1bythefollowingformula:
Takings=2,weseethat(2)isequaltothesumofthereciprocalsofthesquaresofthepositiveintegers:
Convergencecanbeprovenbytheintegraltest,orviathefollowinginequality:
Thisgivesustheupperbound2,andbecausetheinfinitesumhasnonegativeterms,[Link](s)hasa
[Link]=2n:[3]
ArigorousproofusingFourierseries
UseParseval'sidentity(appliedtothefunctionf(x)=x)toobtain
where
forn0,anda0=[Link],
and
Therefore,
asrequired.
Arigorouselementaryproof
Thisisbyfarthemostelementarywellknownproofwhilemostproofsuseresultsfromadvancedmathematics,suchasFourieranalysis,complexanalysis,and
multivariablecalculus,thefollowingdoesnotevenrequiresinglevariablecalculus(althoughasinglelimitistakenattheend).
Foraproofusingtheresiduetheorem,seethelinkedarticle.
Historyofthisproof
TheproofgoesbacktoAugustinLouisCauchy(Coursd'Analyse,1821,NoteVIII).In1954,thisproofappearedinthebookofAkivaandIsaakYaglom
"NonelementaryProblemsinanElementaryExposition".Later,in1982,itappearedinthejournalEureka,attributedtoJohnScholes,butScholesclaimshe
learnedtheprooffromPeterSwinnertonDyer,andinanycasehemaintainstheproofwas"commonknowledgeatCambridgeinthelate1960s".
Theproof
Themainideabehindtheproofistoboundthepartial(finite)sums
2
betweentwoexpressions,eachofwhichwilltendto 6 [Link]
[Link]'sformula,
andwenowturntoestablishingtheseidentities.
Letxbearealnumberwith0<x< ,[Link]'sformulaandthe
2
definitionofthecotangentfunction,wehave
Theinequality
[Link]
squaringgives
Fromthebinomialtheorem,wehave .
(Herecotnxisshorthandfor(cotx)n,andsimilarlyforothertrigonometricfunctions.)
Combiningthetwoequationsandequatingimaginarypartsgivestheidentity
r
Wetakethisidentity,fixapositiveintegerm,setn=2m+1,andconsiderxr=
2m+1 forr=1,2,,[Link]
sin(nxr)=[Link],
foreveryr=1,2,,[Link]=x1,x2,,xmaredistinctnumbersintheinterval0<xr< 2 .Sincethefunctioncot2xisonetooneonthis
interval,thenumberstr=cot2xraredistinctforr=1,2,,[Link],thesemnumbersaretherootsofthemthdegreepolynomial
ByVieta'sformulaswecancalculatethesumoftherootsdirectlybyexaminingthefirsttwocoefficientsofthepolynomial,andthiscomparisonshowsthat
Substitutingtheidentitycsc2x=cot2x+1,wehave
1 r
Nowconsidertheinequalitycot2x< x2 <csc2x(illustratedgeometricallyabove).Ifweaddupalltheseinequalitiesforeachofthenumbersxr= 2m+1 ,and
ifweusethetwoidentitiesabove,weget
2
(
)
Multiplyingthroughby 2m+1 ,thisbecomes
2
Asmapproachesinfinity,theleftandrighthandexpressionseachapproach 6 ,sobythesqueezetheorem,
andthiscompletestheproof.
Seealso
Riemannzetafunction
Apry'sconstant
References
Weil,Andr(1983),NumberTheory:AnApproachThroughHistory,SpringerVerlag,ISBN0817631410.
Dunham,William(1999),Euler:TheMasterofUsAll,MathematicalAssociationofAmerica,ISBN0883853280.
Derbyshire,John(2003),PrimeObsession:BernhardRiemannandtheGreatestUnsolvedProbleminMathematics,JosephHenryPress,ISBN0309
085497.
Aigner,MartinZiegler,GnterM.(1998),ProofsfromTHEBOOK,Berlin,NewYork:SpringerVerlag
Edwards,HaroldM.(2001),Riemann'sZetaFunction,Dover,ISBN0486417409.
Notes
[Link],Raymond(1974)."Eulerandthezetafunction".[Link].81:[Link].2307/2319041.
2.E41Desummisserierumreciprocarum([Link]
[Link],TsuneoIbukiyama,TomoyoshiKaneko,Masanobu(2014).[Link].p.61.
Externallinks
Aninfiniteseriesofsurprises([Link]
"Remarquessurunbeaurapportentrelesseriesdespuissancestantdirectesquereciproques"(PDF).(348kB),EnglishtranslationwithnotesofEulerspaper
[Link]
"HowEulerdidit"(PDF).(265kB)
"TheinfiniteseriesofEulerandtheBernoulli'sspiceupacalculusclass"(PDF).(106kB)
"Evaluating(2)"(PDF).(184kB),FourteenproofscompiledbyRobinChapman
VisualizationofEuler'sfactorizationofthesinefunction([Link]
Retrievedfrom"[Link]
Categories: Numbertheory ZetaandLfunctions Mathematicalproblems
Thispagewaslastmodifiedon17March2017,at02:07.
[Link],youagreetotheTermsofUse
[Link],Inc.,anonprofitorganization.