Jacobianmatrixanddeterminant

FromWikipedia,thefreeencyclopedia

Invectorcalculus,theJacobianmatrix(/dkobin/,/jkobin/)isthematrixofallfirstorderpartial
derivativesofavectorvaluedfunction.Whenthematrixisasquarematrix,boththematrixandits
determinantarereferredtoastheJacobianinliterature.[1]
Supposef:nmisafunctionwhichtakesasinputthevectorxnandproducesasoutputthe
vectorf(x)m.ThentheJacobianmatrixJoffisanmnmatrix,usuallydefinedandarrangedas
follows:

or,componentwise:

(f ,...,f )

Thismatrix,whoseentriesarefunctionsofx,isalsodenotedbyDf,Jf,and (x1 ,...,xm) .(Notethatsome

literaturedefinestheJacobianasthetransposeofthematrixgivenabove.)
TheJacobianmatrixisimportantbecauseifthefunctionfisdifferentiableatapointx(thisisaslightly
strongerconditionthanmerelyrequiringthatallpartialderivativesexistthere),thentheJacobianmatrix
definesalinearmapnm,whichisthebestlinearapproximationofthefunctionfnearthepointx.
Thislinearmapisthusthegeneralizationoftheusualnotionofderivative,andiscalledthederivativeor
thedifferentialoffatx.
Ifm=n,theJacobianmatrixisasquarematrix,anditsdeterminant,afunctionofx1,,xn,isthe
Jacobiandeterminantoff.Itcarriesimportantinformationaboutthelocalbehavioroff.Inparticular,the
functionfhaslocallyintheneighborhoodofapointxaninversefunctionthatisdifferentiableifandonly
iftheJacobiandeterminantisnonzeroatx(seeJacobianconjecture).TheJacobiandeterminantoccursalso
whenchangingthevariablesinmultivariableintegrals(seesubstitutionruleformultiplevariables).
Ifm=1,fisascalarfieldandtheJacobianmatrixisreducedtoarowvectorofpartialderivativesoff
i.e.thegradientoff.
TheseconceptsarenamedafterthemathematicianCarlGustavJacobJacobi(18041851).

Contents

Contents
1 Jacobianmatrix
2 Jacobiandeterminant
3 Inverse
4 Criticalpoints
5 Examples
5.1 Example1
5.2 Example2:polarCartesiantransformation
5.3 Example3:sphericalCartesiantransformation
5.4 Example4
5.5 Example5
6 Otheruses
6.1 Dynamicalsystems
6.2 Newton'smethod
7 Seealso
8 References
9 Furtherreading
10 Externallinks

Jacobianmatrix
TheJacobiangeneralizesthegradientofascalarvaluedfunctionofmultiplevariables,whichitself
generalizesthederivativeofascalarvaluedfunctionofasinglevariable.Inotherwords,theJacobianfora
scalarvaluedmultivariablefunctionisthegradientandthatofascalarvaluedfunctionofsinglevariableis
simplyitsderivative.TheJacobiancanalsobethoughtofasdescribingtheamountof"stretching",
"rotating"or"transforming"thatatransformationimposeslocally.Forexample,if(x,y)=f(x,y)isused
totransformanimage,theJacobianJf(x,y),describeshowtheimageintheneighborhoodof(x,y)is
transformed.

Ifafunctionisdifferentiableatapoint,itsderivativeisgivenincoordinatesbytheJacobian,butafunction
doesn'tneedtobedifferentiablefortheJacobiantobedefined,sinceonlythepartialderivativesare
requiredtoexist.
Ifpisapointinnandfisdifferentiableatp,thenitsderivativeisgivenbyJf(p).Inthiscase,thelinear
mapdescribedbyJf(p)isthebestlinearapproximationoffnearthepointp,inthesensethat

forxclosetopandwhereoisthelittleonotation(forxp)andxpisthedistancebetweenxand
p.
ComparethistoaTaylorseriesforascalarfunctionofascalarargument,truncatedtofirstorder:

Inasense,boththegradientandJacobianare"firstderivatives"theformerthefirstderivativeofascalar
functionofseveralvariables,thelatterthefirstderivativeofavectorfunctionofseveralvariables.
TheJacobianofthegradientofascalarfunctionofseveralvariableshasaspecialname:theHessian
matrix,whichinasenseisthe"secondderivative"ofthefunctioninquestion.

Jacobiandeterminant
Ifm=n,thenfisafunctionfromnto
itselfandtheJacobianmatrixisasquare
matrix.Wecanthenformits
determinant,knownastheJacobian
determinant.TheJacobiandeterminant
isoccasionallyreferredtoas"the
Jacobian".
TheJacobiandeterminantatagiven
pointgivesimportantinformationabout
thebehavioroffnearthatpoint.For
Anonlinearmapf:R2R2sendsasmallsquaretoadistorted
instance,thecontinuouslydifferentiable
parallelepipedclosetotheimageofthesquareunderthebestlinear
functionfisinvertiblenearapoint
approximationoffnearthepoint.
pniftheJacobiandeterminantat
pisnonzero.Thisistheinverse
functiontheorem.Furthermore,iftheJacobiandeterminantatpispositive,thenfpreservesorientation
nearpifitisnegative,freversesorientation.TheabsolutevalueoftheJacobiandeterminantatpgivesus
thefactorbywhichthefunctionfexpandsorshrinksvolumesnearpthisiswhyitoccursinthegeneral
substitutionrule.
TheJacobiandeterminantisusedwhenmakingachangeofvariableswhenevaluatingamultipleintegralof
afunctionoveraregionwithinitsdomain.Toaccommodateforthechangeofcoordinatesthemagnitudeof
theJacobiandeterminantarisesasamultiplicativefactorwithintheintegral.Thisisbecausethen
dimensionaldVelementisingeneralaparallelepipedinthenewcoordinatesystem,andthenvolumeofa

dimensionaldVelementisingeneralaparallelepipedinthenewcoordinatesystem,andthenvolumeofa
parallelepipedisthedeterminantofitsedgevectors.
TheJacobiancanalsobeusedtosolvesystemsofdifferentialequationsatanequilibriumpointor
approximatesolutionsnearanequilibriumpoint.

Inverse
Accordingtotheinversefunctiontheorem,thematrixinverseoftheJacobianmatrixofaninvertible
functionistheJacobianmatrixoftheinversefunction.Thatis,iftheJacobianofthefunction
f:nniscontinuousandnonsingularatthepointpinn,thenfisinvertiblewhenrestrictedto
someneighborhoodofpand

Conversely,iftheJacobiandeterminantisnotzeroatapoint,thenthefunctionislocallyinvertiblenearthis
point,thatisthereisneighbourhoodofthispoint,inwhichthefunctionisinvertible.
The(unproved)Jacobianconjectureisrelatedtoglobalinvertibilityinthecaseofapolynomialfunctions,
thatisafunctiondefinedbynpolynomialsinnvariables.Itassertsthat,iftheJacobiandeterminantisa
nonzeroconstant(or,equivalently,thatitdoesnothaveanycomplexzero),thenthefunctionisinvertible
anditsinverseisapolynomialfunction.

Criticalpoints
Iff:nmisadifferentiablefunction,acriticalpointoffisapointwheretherankoftheJacobian
matrixisnotmaximal.Thismeansthattherankatthecriticalpointislowerthantherankatsome
neighbourpoint.Inotherwords,letkbethemaximaldimensionoftheopenballscontainedintheimageof
fthenapointiscriticalifallminorsofrankkoffarezero.
Inthecasewhere1=m=n=k,apointiscriticaliftheJacobiandeterminantiszero.

Examples
Example1
Considerthefunctionf:22givenby

Thenwehave

and

andtheJacobianmatrixofFis

andtheJacobiandeterminantis

Example2:polarCartesiantransformation
Thetransformationfrompolarcoordinates(r,)toCartesiancoordinates(x,y),isgivenbythefunction
F:+[0,2)2withcomponents:

TheJacobiandeterminantisequaltor.Thiscanbeusedtotransformintegralsbetweenthetwocoordinate
systems:

Example3:sphericalCartesiantransformation
Thetransformationfromsphericalcoordinates(r,,)toCartesiancoordinates(x,y,z),isgivenbythe
functionF:+ [0,][0,2)3withcomponents:

TheJacobianmatrixforthiscoordinatechangeis

Thedeterminantisr2sin.Asanexample,sincedV=dx1dx2dx3thisdeterminantimpliesthatthe
differentialvolumeelementdV=r2sindrdd.Neverthelessthisdeterminantvarieswith
coordinates.

Example4
TheJacobianmatrixofthefunctionF:34withcomponents

is

ThisexampleshowsthattheJacobianneednotbeasquarematrix.

Example5
TheJacobiandeterminantofthefunctionF:33withcomponents

is

FromthisweseethatFreversesorientationnearthosepointswherex1andx2havethesamesignthe
functionislocallyinvertibleeverywhereexceptnearpointswherex1=0orx2=0.Intuitively,ifone
startswithatinyobjectaroundthepoint(1,2,3)andapplyFtothatobject,onewillgetaresultingobject
withapproximately4012=80timesthevolumeoftheoriginalone.

Otheruses
TheJacobianservesasalinearizeddesignmatrixinstatisticalregressionandcurvefittingseenonlinear
leastsquares.

Dynamicalsystems
Consideradynamicalsystemoftheformx=F(x),wherexisthe(componentwise)timederivativeof
x,andF:nnisdifferentiable.IfF(x0)=0,thenx0isastationarypoint(alsocalledacritical
pointthisisnottobeconfusedwithfixedpoints).Thebehaviorofthesystemnearastationarypointis
relatedtotheeigenvaluesofJF(x0),theJacobianofFatthestationarypoint.[2]Specifically,ifthe
eigenvaluesallhaverealpartsthatarenegative,thenthesystemisstablenearthestationarypoint,ifany
eigenvaluehasarealpartthatispositive,thenthepointisunstable.Ifthelargestrealpartofthe
eigenvaluesiszero,theJacobianmatrixdoesnotallowforanevaluationofthestability.

Newton'smethod
AsystemofcouplednonlinearequationscanbesolvediterativelybyNewton'smethod.Thismethoduses
theJacobianmatrixofthesystemofequations.

Seealso
Hessianmatrix
Pushforward(differential)

References
1. Mathworld(http://mathworld.wolfram.com/Jacobian.html)
2. Arrowsmith,D.K.Place,C.M.(1992)."Section3.3".DynamicalSystems.London:Chapman&Hall.ISBN0
412390809.

Furtherreading

Gandolfo,Giancarlo(1996).EconomicDynamics(Thirded.).Berlin:Springer.pp.305330.ISBN3
540609881.

Externallinks
Hazewinkel,Michiel,ed.(2001),"Jacobian",EncyclopediaofMathematics,Springer,ISBN9781
556080104
Mathworld(http://mathworld.wolfram.com/Jacobian.html)Amoretechnicalexplanationof
Jacobians
Retrievedfrom"https://en.wikipedia.org/w/index.php?
title=Jacobian_matrix_and_determinant&oldid=688832060"
Categories: Multivariablecalculus Differentialcalculus Generalizationsofthederivative
Determinants Matrices
Thispagewaslastmodifiedon3November2015,at09:17.
TextisavailableundertheCreativeCommonsAttributionShareAlikeLicenseadditionaltermsmay
apply.Byusingthissite,youagreetotheTermsofUseandPrivacyPolicy.Wikipediaisa
registeredtrademarkoftheWikimediaFoundation,Inc.,anonprofitorganization.