Page - 125 - in Differential Geometrical Theory of Statistics
Image of the Page - 125 -
Text of the Page - 125 -
Entropy2016,18, 254
whereψ → ψ′= a ·ψ is the induced action of the one ofG onN . It isworth observing it is just (6)with
μ= ψ(η) andμ′= ψ′(η). In this sense, the values of themomentummap are just the components of the
momentumG-tensorsdefined in thepreviousSection.
Remark3. WesawatRemarkofSection3thatthemomentumG-tensorμ is identifiedtotheorbitμ= orb(μ, f)
and,disregarding the frames for simplification,wecan identifyμ to the componentorbit orb(μ).
5. LieGroupStatisticalMechanics
In order to discover the underlined geometric structure of the statisticalmechanics, we are
interested in theaffinemapsΘon theaffine spaceofmomentumtensors. Inanaffine frame,Θ is
representedbyanaffinefunctionΘ fromg∗ intoR:
Θ(μ)=Θ(μ)= z+μZ ,
where z=Θ(0)=Θ(μ0)andZ= lin(Θ)∈ gare theaffinecomponentsofΘ. If thecomponentsof
themomentumtensorsaremodifiedaccordingto (6), thechangeofaffinecomponentsofΘ isgivenby
the inducedaction:
z= z′−θ(a)Ad(a)Z′, Z=Ad(a)Z′ . (9)
ThenΘ isaG-tensors. In [3,4],Souriauproposedastatisticalmechanicsmodelusinggeometric
tools. Letdλbeameasureonμ= orb(μ)andaGibbsprobabilitymeasure pdλwith:
p= e−Θ(μ) = e−(z+μZ) .
Thenormalizationcondition ∫
orb(μ)pdλ=1 links thecomponentsofΘ, allowingtoexpressz in
termsofZ:
z(Z)= ln ∫
orb(μ) e−μZdλ . (10)
Thecorrespondingentropyandmeanmomentaare:
s(Z) = − ∫
orb(μ) p lnpdλ= z+MZ,
M(Z) = ∫
orb(μ) μpdλ=−∂z
∂Z , (11)
satisfyingthesametransformationlawastheone(6)ofμ.HenceMarethecomponentsofamomentum
tensorMwhichcanbeidentifiedtotheorbitorb(M), thatdefinesamapμ →M, i.e., acorrespondance
betweentwoorbits. Thisconstruction is formaland, forreasonsof integrability, the integralswillbe
performedonlyonasubsetof theorbitaccordingtoanheuristicwayexplainedlatteron.
Peoplegenerallyconsider that thedefinitionof theentropy is relevant forapplications insofaras
thenumberofparticles in thesystemisveryhuge. For instance, thenumberofatomscontained inone
mole isAvogadro’snumberequal to6×1023. It isworthnotingthatValléeandLerintiuproposeda
generalizationof the idealgas lawbasedonconvexanalysisandadefinitionofentropywhichdoes
not require theclassicalapproximations (Stirling’sFormula) [11].
6.RelativisticThermodynamicsofContinua
Independently of his statistical mechanics, Souriau proposed in [7,8] a thermodynamics of
continua compatiblewith general relativity. Following in his footsteps, one can quote theworks
by Iglesias [12]andVallée [13]. InhisPh.Dthesis,Vallée studied the invariant formofconstitutive
laws in thecontextof special relativitywhere thegravitationeffectsareneglected. In [14], theauthor
andValléeproposedaGalileanversionof this theoryofwhichwerecall thecornerstoneresults. For
moredetails, thereader is referredto [2].
125
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