Page - 122 - in Differential Geometrical Theory of Statistics
Image of the Page - 122 -
Text of the Page - 122 -
Entropy2016,18, 254
brieflythecornerstoneresultsof theGalileanversionofa thermodynamicsofcontinuacompatible
withgeneral relativityproposedbySouriau in [7,8] independently of his statisticalmechanics. In
Section7,wereveal the linkbetweenthepreviousrelativistic thermodynamicsofcontinuaandLie
groupstatisticalmechanics in theclassicalGalileancontext,working insevensteps.
2.AffineTensors
Pointsofanaffinespace.LetAT beanaffinespaceassociatedtoalinearspaceT offinitedimension
n. By thechoiceofanaffineframe f composedofabasisofT andanorigina0,wecanassociate to
eachpointaasetofn (affine)componentsVigathered in then-columnV∈Rn. Forachangeofaffine
frames, the transformation lawfor thecomponentsofapoint reads:
V=C+PV′ , (1)
which isanaffinerepresentationof theaffinegroupofRndenotedAf f(n). It is clearlydifferent from
theusual transformation lawofvectorsV=PV′.
Affine forms.TheaffinemapsΨ fromAT intoRarecalledaffineformsandtheirset isdenoted
A∗T . Inanaffineframe,Ψ is representedbyanaffinefunctionΨ fromRn intoR.Hence, itholds:
Ψ(a)=Ψ(V)=χ+ΦV ,
whereχ=Ψ(0)=Ψ(a0)andΦ= lin(Ψ) is an-row. WecallΦ1,Φ2, · · · ,Φn,χ the componentsof
Ψor, equivalently, the coupleofχ and the rowΦ collecting theΦα. The setA∗T is a linear space
ofdimension (n+1) called thevectordualofAT . Ifwechange theaffine frame, thecomponents
ofanaffine formaremodifiedaccording to the inducedactionofAf f(n), that leads to, taking into
account (1):
χ′=χ−ΦP−1C, Φ′=ΦP−1 , (2)
which isa linearrepresentationofAf f(n).
Affine tensors.Wecangeneralize thisconstructionanddefineanaffinetensorasanobject:
• thatassignsasetofcomponents toeachaffineframe f ofanaffinespaceAT offinitedimensionn,
• witha transformation law,whenchangingof frames,which isanaffineora linear representation
ofAf f(n).
With thisdefinition, theaffine tensorsareanaturalgeneralizationof theclassical tensors thatwe
shall call linear tensors, these lastonesbeingtrivialaffinetensors forwhichtheaffinetransformation
a=(C,P)acts throughits linearpartP= lin(a). Anaffinetensorcanbeconstructedasamapwhich
isaffineor linearwithrespect toeachof itsarguments. Similar to the linear tensors, theaffineonescan
beclassified in three families: covariants, contravariantandmixed. Themostsimpleaffine tensorsare
thepointswhichare1-contravariantandtheaffineformswhichare1-covariantbutwecanconstruct
morecomplexoneshavingastrongphysicalmeaning: the torsors (proposedin[5]), the co-torsorsand
themomentaextensivelydetailledin[2]. Formoredetailsontheaffinedualspace,affinetensorproduct,
affinewedgeproductandaffinetangentbundles, thereader interested in this topic is referredto the
so-calledAV-differentialgeometry [9].
G-tensors.AsubgroupGofAf f(n)naturallyactsonto theaffine tensorsbyrestriction toGof
their transformation law.LetFGbeasetofaffineframesofwhichG isa transformationgroup.The
elementsofFG arecalledG-frames.AG-tensor isanobject:
• thatassignsasetofcomponents toeachG-frame f,
• witha transformation law,whenchangingof frames,which isanaffineora linear representation
ofG.
122
Differential Geometrical Theory of Statistics
- Title
- Differential Geometrical Theory of Statistics
- Authors
- Frédéric Barbaresco
- Frank Nielsen
- Editor
- MDPI
- Location
- Basel
- Date
- 2017
- Language
- English
- License
- CC BY-NC-ND 4.0
- ISBN
- 978-3-03842-425-3
- Size
- 17.0 x 24.4 cm
- Pages
- 476
- Keywords
- Entropy, Coding Theory, Maximum entropy, Information geometry, Computational Information Geometry, Hessian Geometry, Divergence Geometry, Information topology, Cohomology, Shape Space, Statistical physics, Thermodynamics
- Categories
- Naturwissenschaften Physik