expansionsofmanytranscendentalfunctionsItalsoplaysanimportantroleincombinatoricsSection214 BecausetheytooariseinthecontextofcombinatoricsStirlingnumbersofthesecondkindarediscussedinSection 214 ID: 267563
Download Pdf The PPT/PDF document "Thefactorialfunctionoccurswidelyinfuncti..." is the property of its rightful owner. Permission is granted to download and print the materials on this web site for personal, non-commercial use only, and to display it on your personal computer provided you do not modify the materials and that you retain all copyright notices contained in the materials. By downloading content from our website, you accept the terms of this agreement.
Thefactorialfunctionoccurswidelyinfunctiontheory;especiallyinthedenominatorsofpowerseries expansionsofmanytranscendentalfunctions.Italsoplaysanimportantroleincombinatorics[Section2:14]. Becausetheytooariseinthecontextofcombinatorics,StirlingnumbersofthesecondkindarediscussedinSection 2:14[thoseofthefirstkindfindahomeinChapter18]. 2:6THEFACTORIALFUNCTION n ! 23 Setting m n inequation2:5:2providesa duplicationformula forthefactorialfunction,enabling(2 n )!tobe expressedwiththehelpofaPochhammerpolynomial.Alternativeduplicationformulasareavailablefromequations 2:12:3and2:12:4,whichmayberewrittenas 1 1 2 2 /2 11 1 2 22 (1)/2 2!2,4,6, ! 2!1,3,5, n n n n nn n nn ° ® ° ¯ 2:5:4 Thereareanalogoustriplicationformulasthatcanbedevelopedfromtheequationsin2:12:5. Thefrequentoccurrenceoffactorialsascoefficientsofpowerseriespermitsthesummationofsuchseriesas 0 1111 exp(1)271828182845905 0!1!2!! j j f ¦ . 2:5:5 and 0 2222 0 1111 I(2)227958530233607 0!1!2!! j j f ¦ . 2:5:6 whereI 0 isthemodifiedBesselfunction[Chapter49].Thecorrespondingserieswithalternatingsignssumsimilarly toexp( 1)andtotheparticularvalueJ 0 (2)ofthezero-orderBesselfunction[Chapter52].Thereiseventhe intriguingasymptoticresult[seeequation37:13:4] 0 0 exp() 0!1!2!()!d0596347362323194 1 j j t jt t f f ¦ ³ . ~ 2:5:7 Moreover,theseries ( 1) n /(2 n )!sumstocos(1)andthereareseveralanalogoussummations. 2:6EXPANSIONS Stirlingsformula [seealsoSection43:6] 23 11139 !2exp()1 1228851840 n nnnnn nnn ªº Sof «» ¬¼ ~ 2:6:1 providesanexpansionforthefactorialfunction.Thoughtheyaretechnicallyasymptotic[Section0:6],this expansionandasimilaroneforthelogarithmofthefactorialfunction 1 35 B 1 ln(!)ln2ln()1 (1) ln(2)111 ln() 2123601260 j j j nnnn jjn n nnnn nnn ªº §· S «» ¨¸ ©¹ «» ¬¼ S of ¦ ~ 2:6:2 areremarkablyaccurate,evenforsmall n .B j isthe j th Bernoullinumber[Chapter4]. 2:7PARTICULARVALUES 0!1!2!3!4!5!6!7!8!9!10!11!12! 112624120720 504040320362880362880039916800479001600 24 THEFACTORIALFUNCTION n !2:8 2:8NUMERICALVALUES Thedecimalintegerrepresenting n !hasexactlyInt( n /5)+Int( n /25)+Int( n /125)+ ### terminalzeros;forexample 31!endswithsevenzeros.Thisruleisusefulincalculatingexactnumericalvaluesoflargefactorials.Intisthe integer-valuefunctiondescribeinChapter8. Equator s factorialfunction routine(keyword ! )providesvaluesof n !.Forintegerinputintherange0 n d 170,asimplealgorithmbasedonrecursion2:5:1,followedbyrounding,isusedtocompute n !.Exactoutputis reportedupto20! 2 . 43290200817664E+18.For21 d n d 170, Equator providesafloatingpointapproximation of n !preciseto15digits. Separately,valuesofthenaturalanddecadiclogarithms,ln( n !)andlog 10 ( n !),areprovidedbythe logarithmic factorialfunction and logarithmtobase10ofthefactorialfunction routines( ln! and log10! ).Suchlogarithmic valuesareusefulwhen n islargebecause n !itselfisthenprohibitivelyhuge.Forinputupto n 170,ln( n !)is computedbysimplytakingthelogarithmoftheoutputfromtheroutinedescribedabove.Forintegerinputinthe range171 d n d 1E305, Equator usesStirlingsformulainthetruncatedandconcatenatedform 222 ln(2)112 ln(!)1ln()11 212307 n nnn nnn §· S§· §· ¨¸ ¨¸ ¨¸ ©¹ ©¹ ©¹ 2:8:1 Divisionby2 . 30258509299405generateslog 10 ( n !). 2:9LIMITSANDAPPROXIMATIONS Forlargeargument,thelimitingformula !2 n n nnn e §· oSof ¨¸ ©¹ 2:9:1 applies, e beingthebaseofnaturallogarithms[Section1:7].Therelatedapproximation !Round(112) 72 n n nn ne ½ S °° §· | ®¾ ¨¸ ©¹ °° ¯¿ 2:9:2 issurprisinglygood,andevenexact,forsmallpositiveintegers.Roundistheroundingfunction,describedin Section8:13. 2:10OPERATIONSOFTHECALCULUS Nooperationsofthecalculusarepossibleonafunctionsuchas n !thatisdefinedonlyfordiscretearguments. 2:11COMPLEXARGUMENT Inviewofrelation2:12:1,thegammafunctionformulasgiveninSection43:11maybeusedtoascribemeaning to( n + im )!. 2:12THEFACTORIALFUNCTION n ! 25 2:12GENERALIZATIONS Thefactorialfunctionisaspecialcaseofthegammafunction[Chapter43] !(1)0,1,2, nnn * 2:12:1 andofthePochhammerpolynomial[Chapter18] !(1)0,1,2, n nn 2:12:2 Thelatteridentitypermitsustowrite 1 2 (2)!4()(1) n nn n 2:12:3 and 3 2 (21)!4(1)() n nn n 2:12:4 Similarly 333 5 12244 333333 (3)!3(1),(31)!3(1),(32)!3(1) nnn nnn nnnnnn nnn 2:12:5 andsoon. 2:13COGNATEFUNCTIONS:multiplefactorials See6:3:4forthecloserelationshipbetweenthefactorialfunctionandbinomialcoefficients. The doublefactorial or semifactorialfunction isdefinedby 11,0 !!(2)(4)5311,3,5, (2)(4)6422,4,6, n nnnnn nnnn ° uuuuuu ® ° uuuuuu ¯ 2:13:1 Forevenargumentitreducesto 2 !!2/2!0,2,4, n nnn 2:13:2 whileforodd n itmaybeexpressedintermsoffactorials,orasagammafunction[Chapter43]orasaPochhammer polynomial[Chapter18] 2 (1)/2 (1)/2 (1)/2 !21 !!2121,3,5, 1 22 2! 2 n n n n nn nn n §·§· * ¨¸¨¸ §·S ©¹©¹ ¨¸ ©¹ 2:13:3 Equivalenttothelastequationis (21)! (21)!! 2! n n n n 2:13:4 Theseformulasareusedby Equator s doublefactorialfunction routine(keyword !! )tocomputevaluesof n !!for integersintherange 1 d n d 300.Ofcourse !!(1)!!!0,1,2, nnnn 2:13:5 Someearlymembersofthedouble-factorialfamilyarelistedbelow. ( 1)!!0!!1!!2!!3!!4!!5!!6!!7!!8!!9!!10!!11!!12!!13!!14!! 1112381548105384945 38401039546080135135645120 26 THEFACTORIALFUNCTION n !2:14 Notethat,apartfrom0!! 1,thedoublefactorial n !!sharestheparityof n .Alsonoticethat,toaccordwiththe n 1instanceofthegeneralrecursionformula (2)!!(2)!! nnn 2:13:6 ( 1)!!isassignedthevalueofunity.Withasimilarrationale,onesometimesencountersthevalues( 3)!! 1, ( 5)!! ,etc. 1 3 Offrequentoccurrence[forexampleinSections6:4,32:5,61:6and62:12]istheratio( n 1)!!/ n !!ofthedouble factorialsofconsecutiveintegers.Forodd n ,theratioisexpressiblebytheintegral 2 /2 1 0 (1)!!21 !sin()d1,3,5, !!!2 n n nn ttn nn S ªº §· ¨¸ «» ©¹ ¬¼ ³ 2:13:7 whileforeven n itisgivenby Wallissformula (JohnWallis,Englishmathematicianandcryptographer,1616 1703) @ /2 2 0 (1)!!!2 sin()d0,2,4, !! 2(/2)! n n nn ttn n n S S ³ 2:13:8 Thisimportantratiohastheasymptoticexpansion 2 2 211 1even 432 (1)!! !! 11 1odd 2432 n nnn n n n nnn ªº of ° «» S °¬¼ ® S ªº ° of «» ° ¬¼ ¯ ~ 2:13:9 Finitesumsofsomesuchratiosobeythesimplerule 0 1315(21)!!(21)!!(21)!! 10,1,2, 2848(2)!!(2)!!(2)!! n j njn n njn ¦ 2:13:10 andthereistherelatedinfinitesummationduetoRoss: 1 1315105(21)!! ln(4) 2161441536(2)!! j j jj f ¦ 2:13:11 The triplefactorial isdefinedanalogously 12,1,0 (3)(6)7411,4,7, !!! (3)(6)8522,5,8, (3)(6)9633,6,9, n nnnn n nnnn nnnn ° uuuuuu ° ® uuuuuu ° ° uuuuuu ¯ 2:13:12 andfindsapplicationinconnectionswithAiryfunctions[Chapter56].Someearlyvaluesare: ( 2)!!!( 1)!!!0!!!1!!!2!!!3!!!4!!!5!!!6!!!7!!!8!!!9!!!10!!!11!!!12!!! 111123410182880162280880 1944 Theextensiontoa quadruplefactorialn !!!!isobvious;itisusefulinSections43:4and59:7. 2:14RELATEDTOPIC:combinatoricsandStirlingnumbersofthesecondkind Thefactorialfunctionappearsveryofteninapplicationsinvolving combinatorics .Forexample,thenumber 2:14THEFACTORIALFUNCTION n ! 27 of permutations (arrangements)of n objects,alldifferent,is n !.Ifnotallofthe n objectsaredifferent,thenumber ofpermutationsisreducedto 12 12 ! ()!()!()! J J n nnnn nnn 2:14:1 wherethereare n 1 samplesofobject1, n 2 samplesofobject2, ### , n J samplesofobject J ,sothat n j n . Iffromagroupof n objects,alldifferent,onewithdraws m objects,oneatatime,thenumberof variations (possiblewithdrawalsequences)is ! ()! n mn nm d 2:14:2 Ifoneignorestheorderofwithdrawal,2:14:2isreducedto ! ()!! n mn nmm d 2:14:3 whichthenrepresentsthenumberofwaysinwhich m objectscanbechosenfromamong n ,alldifferent,andis knownasthenumberof combinations .Expression2:14:3is,infact,thebinomialcoefficientaddressedinChapter6. Thenumberof partitions (differentwaysinwhich n distinctobjectsmaybeplacedin m identicalboxessothat eachboxcontainsatleastoneobject)isgivenbya Stirlingnumberofthesecondkind .Thereisnostandardized notationforsuchfunctions;this Atlas usesthesymbol.Clearly,nopartitioningispossibleif m 0orif n m () m n V andaccordinglythesecondStirlingnumberiszeroinsuchcircumstances.Otherwise,ageneralformulais () 0 () ()!! mjn m m n j j mjj V ¦ 2:14:4 Non-zerovaluesofStirlingnumbersof thesecondkindare,ofcourse,positive integersandsomeareshownin Figure2-2.Othersmaybecalculated viatherecursionformula ()()(1) 11 mmm nnn m V VV 2:14:5 for m 1,2,3, ### and n 1,2,3, ### .This recursionformsthebasisof Equator s Stirlingnumberofthesecondkind routine(keyword sigmanum ).First, andareinitializedtozero (1) 0 m o V (0) 1 n o V whileissetequalto1.Then (0) 0 V ,, ### ,arecalculated (1) 1 m o V (1) 2 m o V (1) m n o V viarecursion2:14:5.BecauseStirling numbersareintegers,roundingensures thatthe15digitsthat Equator generates areexact.