1. PHYSICAL REVIEW D 81, 065013 (2010)
ˇ
Conformal anisotropic mechanics and the Horava dispersion relation
Juan M. Romero*
´ ´ ´
Departamento de Matematicas Aplicadas y Sistemas, Universidad Autonoma Metropolitana-Cuajimalpa, Mexico 01120 DF, Mexico
Vladimir Cuesta,† J. Antonio Garcia,‡ and J. David Vergarax
´ ´ ´
Instituto de Ciencias Nucleares, Universidad Nacional Autonoma de Mexico, Apartado Postal 70-543, Mexico 04510 DF, Mexico
(Received 22 September 2009; revised manuscript received 5 February 2010; published 11 March 2010)
In this paper we implement scale anisotropic transformations between space and time in classical
ˇ ˇ
mechanics. The resulting system is consistent with the dispersion relation of Horava gravity [P. Horava,
Phys. Rev. D 79, 084008 (2009)]. Also, we show that our model is a generalization of the conformal
mechanics of Alfaro, Fubini, and Furlan. For an arbitrary dynamical exponent we construct the dynamical
¨
symmetries that correspond to the Schrodinger algebra. Furthermore, we obtain the Boltzmann distribu-
tion for a gas of free particles compatible with anisotropic scaling transformations and compare our result
with the corresponding thermodynamics of the recent anisotropic black branes proposed in the literature.
DOI: 10.1103/PhysRevD.81.065013 PACS numbers: 11.10.Ef, 11.10.Àz, 11.25.Hf
I. INTRODUCTION ghost free. On the other hand due to the nonrelativistic
ˇ
character of the Horava gravity, this theory must be con-
Scaling space-time symmetries are very useful tools to
sidered as an effective version of a more fundamental
study nonlinear physical systems and critical phenomena
theory, compatible with the symmetries that we observe
[1,2]. At the level of classical mechanics an interesting
in our universe. The consideration of these models could be
example is the conformal mechanics [3] that is relevant for
useful to understand the quantum effects of gravity.
several physical systems from molecular physics to black
In order to gain some understanding of the content of
holes [4,5]. Recently a new class of anisotropic scaling
generalized dispersion relations of the form (2), we will
transformations between spatial and time coordinates has
construct a reparametrization invariant classical mechanics
been considered in the context of critical phenomena [6],
compatible with the scaling laws (1). In particular this
string theory [7–9], and quantum gravity [10]:
implementation of the scaling transformations (1) will
t ! bz t; x ! bx;
~ ~ (1) lead to dispersion relations of the form (2). Another inter-
esting characteristic of our implementation is that it con-
where z plays the role of a dynamical critical exponent. In
tains several systems for different limits. In the case of
ˇ
particular the result of Horava [10] is quite interesting
z ¼ 2 the conformal mechanics of Alfaro, Fubini, and
since it produces a theory that is similar to general relativ-
Furlan [3] is recovered. In the limit z ! 1 we get a
ity at large scales and may provide a candidate for a UV
relativistic particle with Euclidean signature, and in the
completion of general relativity. In principle this model for
limit z ¼ 1 together with m ! 0 we obtain a massless
gravity is renormalizable in the sense that the effective
relativistic particle.
coupling constant is dimensionless in the UV limit. A
We also study the symmetries of the anisotropic me-
fundamental property of these gravity models is that the
chanics and compare our results with the Schrodinger¨
scaling laws (1) are not compatible with the usual relativ-
algebra for any value of z. Finally, by analyzing the ther-
ity. In the IR these systems approach the usual general
modynamic properties of a free gas compatible with the
relativity with local Lorentz invariance but in the UV the
scaling transformation (1), we will obtain the same ther-
formulation admits generalized dispersion relations of the
modynamic relations for black branes recently proposed in
form
[7].
P2 À GðP2 Þz ¼ 0;
0
~ G ¼ const: (2) The organization of the paper is as follows: Section II
concerns the action of the system. In Sec. III we study the
Here z takes also the role of a dynamical exponent. As the canonical formalism of the system and its gauge symme-
dispersion relations used by these models are quadratic in tries. Section IV focuses on the equations of motion. The
P0 , while the spatial momentum scales as z, the models are ¨
global symmetries and the Schrodinger algebra are ana-
in principle renormalizable by power counting arguments lyzed in Sec. V. In Sec. VI we consider the thermodynamic
at least for z ¼ 3. It is claimed also that these models are properties of the system and we conclude in Sec. VII.
*jromero@correo.cua.uam.mx II. THE ACTION PRINCIPLE
†
vladimir.cuesta@nucleares.unam.mx
‡
garcia@nucleares.unam.mx Scale transformations can be realized in nonrelativistic
x
vergara@nucleares.unam.mx theories in different ways. On the one hand we can start
1550-7998= 2010=81(6)=065013(7) 065013-1 Ó 2010 The American Physical Society

2. ROMERO et al. PHYSICAL REVIEW D 81, 065013 (2010)
Z 2 z=2
from an action in curved space where some of its isome- m ðx Þ
_
tries are the scale transformations. Another way to imple- S¼ d þ ðz À 2Þe : (9)
2ðz À 1Þ ezÀ2 t_
ment scale transformations is to start from an action in flat
space but constructed in such a way that the invariance Here, we have introduced an einbein e in a similar way as
under scale transformations is manifest. In this work we the standard trick to remove the square root in the action of
will take this second point of view. Let us consider a d þ the free relativistic particle. Using this action (9) we can
1-dimensional space-time, with an action principle given now take the limit z ! 1 if at the same time we implement
by the limit m ! 0. We assume that m ! 0 at the same rate as
z À 1 in such a way that zÀ1 ! , where plays the role
m
Z Z m ðx2 Þz=2ðzÀ1Þ
_
dt
S ¼ dL ¼ d À VðxÞt~ _ ; t_ ¼ ; of a mass parameter. In consequence we obtain the action
2 ðtÞ _ 1=zÀ1 d pffiffiffiffiffi
Z e x2 _
xi ¼
_
dxi
; x2 ¼ xi xi ;
_ _ _ i ¼ 1; 2; Á Á Á ; d: (3) S ¼ d À1 : (10)
d 2 t_
This action is invariant under reparametrizations of , i.e., This action is still invariant under reparametrizations due
under the transformations ! ¼ ðfÞ, if we require that to the transformation rule of the einbein e ! df e, and it
d
t and the xi are scalars under reparametrizations. corresponds to a relativistic massless particle whose usual
Furthermore, if we consider that the potential satisfies covariant action is
VðbxÞ ¼ bÀz VðxÞ, for example,
~ ~ Z x x
_ _
a S ¼ d : (11)
VðxÞ ¼ z ;
~ a ¼ const; (4) 2e
jxj
~
the action (3) is invariant under the scaling laws (1). Note
III. CANONICAL FORMALISM
that, when z ¼ 2 and ¼ t, we obtain the action for a
nonrelativistic particle; furthermore taking into account Let us consider the canonical formalism of the system
(4), we obtain (3). The canonical momenta are
Z m a
@L mz ðx2 Þ2Àz=2ðzÀ1Þ
_
S ¼ dt x À 2 ;
_ 2
(5) pi ¼ ¼ _
xi ; (12)
2 jxj
~ _
@x i 2ðz À 1Þ ðtÞ1=zÀ1
_
which corresponds to the action of the conformal mechan-
ics [3]. In the case of arbitrary z and for a gauge condition @L m ðx2 Þz=2ðzÀ1Þ
_
pt ¼ ¼À þ VðxÞ ¼ ÀH;
~
¼ t we get @t_ 2ðz À 1Þ ðtÞz=zÀ1
_
Z m a
(13)
S ¼ dt ðx2 Þz=2ðzÀ1Þ À z ;
_ (6)
2 jxj
~ and as expected (by the reparametrization invariance) the
which again is invariant under the scaling laws (1). In this canonical Hamiltonian vanishes,
sense Eq. (3) represents a generalized conformal Hc ¼ pi xi þ pt t_ À L ¼ 0:
_ (14)
mechanics.
Another interesting property of the action (3) appears in From (12) we get
the limit z ! 1. In this case for an arbitrary potential we
mz ðx2 Þ1=2ðzÀ1Þ 2
_
obtain p2 ¼ pi pi ¼ ; (15)
2ðz À 1Þ ðtÞ1=zÀ1
_
Z m pffiffiffiffiffi
S ¼ d x2 À VðxÞt_ ;
_ ~ (7) which in turn implies
2 Àz
m mz
this action is invariant under reparametrizations. In the ¼ pt þ ðp2 Þz=2 þ VðxÞ % 0:
~
2ðz À 1Þ 2ðz À 1Þ
particular case of the potential (4) the limit is singular in
the region 0 jxj 1; then leaving VðxÞ ¼ 0 we obtain
~ (16)
the action for a relativistic particle with Euclidean signa- In the simple case VðxÞ ¼ 0 when we enforce the con-
~
ture, straint we obtain
Z m pffiffiffiffiffi Àz
S ¼ d x2 : m mz
_ (8) pt ¼ À ðp2 Þz=2 ; (17)
2 2ðz À 1Þ 2ðz À 1Þ
The case z ! 1 is also very interesting. Even though this which can be rewritten in the form
limit is not well defined for the action (3) it can be 2ð1ÀzÞ
implemented using an equivalent form of the action (3). 1 m
ðpt Þ2 À Gðp2 Þz ¼ 0; G¼ : (18)
In the case VðxÞ ¼ 0 we get
~ z2z 2ðz À 1Þ
065013-2

3. CONFORMAL ANISOTROPIC MECHANICS AND THE . . . PHYSICAL REVIEW D 81, 065013 (2010)
pffiffiffiffi
In this way, we get a dispersion relation that is similar to xi ¼ fxi ; g ¼ Gzðp2 ÞzÀ2=2 pi ;
ˇ
that proposed in Horava gravity [Eq. (2)]. The relation (17)
@V
implies that the momentum pt is negative for z 1. In pi ¼ fpi ; g ¼ À ; t ¼ ft; g ¼ ;
order to preserve the equivalence between (17) and (18) for @xi
z 1 we assume that pt takes only negative values in (18). @
Notice that by taking the limits z ! 1 and m ! 0 in (16) pt ¼ 0; ¼ :
@
we get the constraint
Here f; g are the Poisson brackets. The above transforma-
¼ pt þ ðp2 Þ1=2 þ VðxÞ:
~ (19) tions are of the usual type for the parametrized particle.
However, the rule for the transformation corresponding to
In the case VðxÞ ¼ 0 this constraint is consistent with the
the spatial coordinates seems to change drastically. But if
canonical formalism of the action (9) and corresponds to a
we use the definition of the momenta (12) we will get
massless relativistic free particle. i
exactly the usual transformation xi ¼ dx recovering
Now, from the Eq. (16) the system has one first class dt
constraint and, following the standard Dirac’s method [11], the full diffeomorphism transformations associated with
the ‘‘wave equation’’ the reparametrization invariance. This result can be con-
pffiffiffiffi ˇ
trasted with the same result obtained by Horava’s in his
j c ¼ ðpt þ Gðp2 Þz=2 þ VðxÞÞj c ¼ 0
^ ^ ^ ~ (20) gravity model [10].
In this way we saw that the action Eq. (3) has very
must be implemented on the Hilbert space at quantum level ˇ
similar properties to Horava’s gravity, namely, the same
to obtain the physical states. In other words, the scaling properties and the same invariance under repara-
¨
‘‘Schrodinger equation’’ corresponding to the anisotropic metrizations. In this sense, we see that our toy model has
mechanics (3) is ˇ
the same transformation properties as Horava’s gravity,
@c pffiffiffiffi just as the relativistic particle has the same transformation
i@ ¼ ð GðÀ@2 r2 Þz=2 þ VðxÞÞ c :
~ (21) properties as general relativity [17].
@t
¨
It is interesting to notice that this type of ‘‘Schrodinger
IV. EQUATIONS OF MOTION
equation’’ has been already considered in the literature
(see, for example, [12]). The original motivation behind From the variation of the action (3), and taking into
this equation comes from the anomalous diffusion equation account the definitions of the momenta pt and pi , it follows
´ ˇ
and Levy flights. For a perspective from Horava gravity see that
Z d
[13] where a relation with spectral dimension and fractals
was proposed. Furthermore, a proposal for the free particle dpi @V
S ¼ d ðp x þ pt tÞ À xi þ
quantum-mechanical kernel associated to (21) was given in d i i d @xi
terms of the Fox’s H function [14]. On the other hand, in
dp
the limit z ! 1, the Schrodinger equation follows from
¨ þ t t ; (24)
d
the action (7); for arbitrary potential we get a primary
constraint p2 À m ¼ 0 and several secondary constraints.
2
4 where the Euler-Lagrange equations are
In the case V ¼ x2 , a similar system has been considered in
[15]. For V ¼ 0, we only have now the first class constraint dpi @V
þ ¼ 0; (25)
p2 À m ¼ 0, which at the quantum level corresponds to
2
d @xi
4
2
@2 r2 þ m c ¼ 0: (22) d
4 pt ¼ 0: (26)
d
This is the expected Helmholtz equation and may be
Notice that for z ¼ 2 and ¼ t Eq. (25) reduces to the
viewed as the Klein-Gordon equation with a Euclidean
usual Newton’s second law, and that (26) corresponds to
signature metric.
the energy conservation.
Because the action (3) is invariant under reparametriza-
In order to simplify some computations we will use the
tions, there is a local symmetry. The extended canonical
gauge condition ¼ t. This is a good gauge condition in
action for the system is [16]
the sense that f; À tg Þ 0. Under the above assumption
Z and using the definition pi we get the equations of motion
S ¼ dðtpt þ xi pi À Þ
_ _ (23)
d mz @V
ðx2 Þð2ÀzÞ=2ðzÀ1Þ xi þ
_ _ ¼ 0: (27)
with a Lagrange multiplier. dt 2ðz À 1Þ @xi
If we require that is a function of , the extended action
is invariant under the gauge transformations [16] We thus get
065013-3

4. ROMERO et al. PHYSICAL REVIEW D 81, 065013 (2010)
2ðz À 1Þ @V this algebra are the Galilean algebra
gij xj ¼ Àðx2 ÞðzÀ2Þ=2ðzÀ1Þ
€ _ ;
mz @xi fJ ij ; J kl g ¼ ðik J jl þ jl J ik À il J jk À jk J il Þ;
(28)
2 À z xi xj
_ _
gij ¼ ij þ : fJ ij ; Pk g ¼ ðik Pj À jk Pi Þ;
z À 1 x2
_
fJ ij ; K k g ¼ ðik K j À jk Ki Þ; (32)
The matrix gij can be considered as a metric that depends
on the velocities. We see that this metric is homogeneous fH; Ki g ¼ Pi ;
of degree zero in the velocities and in consequence is a fPi ; Kj g ¼ Àij M;
Finsler type metric [18]. This metric has three critical
points: plus dilatations
z ¼ 1; z ¼ 2; z ¼ 1: (29) fD; Pi g ¼ Pi ; fD; Ki g ¼ K i ; fD; Hg ¼ À2H;
For z Þ 1, we obtain the inverse metric (33)
xi xj
_ _ and special conformal transformations
gij ¼ ij À ð2 À zÞ 2 ; (30)
_
x
fD; Cg ¼ À2C; fH; Cg ¼ D; fC; Pi g ¼ Ki :
and it follows that (34)
2ð1 À zÞ 2 ðzÀ2Þ=ð2ðzÀ1ÞÞ _ _
xi xj @V
xi ¼
€ ðx Þ
_ ij þ ðz À 2Þ 2 : It is interesting to notice that H, C, D close themselves in
mz x @xj
_ an SLð2; RÞ subalgebra. Thus the Schrodinger algebra is
¨
(31) the Galilean algebra plus dilatation and special conformal
transformations [24]. M is the center of the algebra.
The limit z ¼ 1 is particularly interesting, since in this ¨
It is also worth noticing that the Schrodinger algebra in d
case the matrix gij is not invertible. In this limit we get the
dimensions, SchrðdÞ, can be embedded into the relativistic
action (7) and in this case the constraint (16) is not valid. conformal algebra in d þ 2 space-time dimensions Oðd þ
Furthermore, from the Hamiltonian analysis of the system 2; 2Þ. This fact is central in the recent literature about the
there appear secondary constraints that transform the geometric content of the new anti–de Sitter/nonrelativistic
model into a system with second class constraints [15]. conformal field theories dualities (for details see [9]).
We are interested in the construction of the explicit
V. DYNAMICAL SYMMETRIES OF THE ¨
generator of the relative Poisson-Lie Schrodinger algebra
SCHRODINGER EQUATION FOR ANY z
¨ for any z. These generators will be the symmetries of the
¨
corresponding anisotropic Schrodinger equation (21) asso-
In the same sense that the Galilei group with no central ciated with the nonrelativistic anisotropic classical me-
charges [19] exhausts all the symmetries of the nonrelativ- chanics (3).
¨
istic free particle in d dimensions, the Schrodinger algebra We will denote the generalization of the Schrodinger¨
¨
is the algebra of symmetries of the Schrodinger equation symmetry algebra for any z as Schrz ðdÞ in d dimensions.
with V ¼ 0 and can be considered as the nonrelativistic The algebra is given by the Galilean algebra (32) plus
limit of the conformal symmetry algebra. Another interest-
ing Galilean conformal algebra can also be constructed as fD; Pi g ¼ Pi ; fD; Ki g ¼ ð1 À zÞKi ;
the nonrelativistic contraction of the full conformal rela- (35)
tivistic algebra [21]. In the special case of d ¼ 2 a Iononu-
¨ ¨ fD; Mg ¼ ð2 À zÞM; fD; Hg ¼ ÀzH:
´
Wigner contraction of the Poincare symmetry of a free Notice that M is not playing the role of the center anymore
anyon theory is reduced to a Galilean symmetry with two unless z ¼ 2 and that C, the generator of the special
central charges (see, for example, [22] and references conformal transformations, is not in the algebra [25]. We
¨
therein). The Schrodinger symmetry as a conformal sym- are not aware of any explicit realization of this algebra in
metry of nonrelativistic particle models was introduced phase space for arbitrary dynamical exponent z. For z ¼ 2
some time ago [20] in a holographic context relating a the phase space realization reads
system in AdSdþ1 with a massless particle in
d-dimensional Minkowski space-time. As isometries of J ij ¼ xi pj À xj pi ; H ¼ Àpt ; Pi ¼ pi ;
certain backgrounds and some applications and examples,
¨
the Schrodinger symmetry was studied in [23]. The gen- K i ¼ ÀMxi þ tpi C ¼ t2 pt þ txi pi À 1Mx2 ;
2
¨
erators of the Schrodinger algebra include temporal trans- D ¼ 2tpt þ xi pi :
lations H, spatial translations Pi , rotations J ij , Galilean
boost Ki , dilatation D (where time and space can dilate Inspired by this construction and the embedding of the
with different factors), special conformal transformation ¨
Schrodinger algebra in d dimensions into the full relativ-
C, and the central charge M. The nonzero commutators of istic conformal algebra in d þ 2 [9], we will display a set of
065013-4

5. CONFORMAL ANISOTROPIC MECHANICS AND THE . . . PHYSICAL REVIEW D 81, 065013 (2010)
¨
dynamical symmetries of the Schrodinger equation for any 1
¼À :
z. ð1 À Þ
The crux of the argument is to allow M to depend on the
magnitude of pi squared Mðp2 Þ. We will make the ansatz So we have the nontrivial dynamical symmetries of the
that M is a homogeneous function of degree in pi ¨
anisotropic Schrodinger equation generated by
@M i D ¼ ztpt þ xi pi ; (40)
p ¼ 2M; Mðp2 Þ ¼ Aðp2 Þ ; (36)
@pi
z2 2 z 1
¨
where A is a constant. Then taking as our Schrodinger C¼ t pt þ txi pi À Mðp2 Þx2 ; (41)
4 2 2
equation (16) with VðxÞ ¼ 0 and for any z, we have
~
that are dilatation and special conformal transformations
p2 and d Galilean boosts
S ¼ pt þ H ¼ pt þ : (37)
2Mðp2 Þ z
Ki ¼ tpi À Mðp2 Þxi : (42)
We will ask for the phase space quantities that commute (in 2
the Poisson sense) with Eq. (37), i.e., A similar result was obtained previously in [6] for a
fO; Sg ¼ 0; (38) differential representation of the algebra (35) using the
concept of fractional derivatives.
over S ¼ 0. Any such phase space quantity will be a Unfortunately the dynamical symmetries given by (40)–
¨
dynamical symmetry of the anisotropic Schrodinger equa- (42), plus J ij , Pi , H, M, do not close in Schrz ðdÞ (35).
tion (17). is fixed in terms of the dynamical exponent z Nevertheless a subalgebra formed by H, C, D indeed
by relating our definition (37) with the first class constraint closes into an SLð2; RÞ sector of the full algebra,
(17) z
fD; Hg ¼ ÀzH; fD; Cg ¼ ÀzC; fH; Cg ¼ D:
2Àz 1 2
¼ ; A ¼ pffiffiffiffi : (39)
2 2 G To these generators we can add the angular momentum J ij
that commutes with them.
J ij , Pi , H are obvious symmetries of the anisotropic
¨
Schrodinger equation. To see what are the analogs of D,
C, and Ki , let us start with D as defined for the z ¼ 2 case VI. THERMODYNAMIC PROPERTIES
but with an anisotropic rescaling between space and time
In the following we shall sketch some of the thermody-
D ¼ tpt þ

6. xi pi : namic properties of the system. We will consider only the
case without potential. In this case the energy of our model
By taking the Poisson bracket with S we have is

7. ð1 À Þp2 pffiffiffiffi
fD; Sg ¼ pt þ ; H ¼ Gpz : (43)
Mðp2 Þ
Let us denote by V the volume and

8. ¼ kT , with T the
1
then temperature and k the Boltzmann constant, that the canoni-
1 cal partition function for the system in a space of d dimen-

9. ¼ : sions is
2ð1 À Þ
V Z pffiffiffi z V Z pffiffiffi z
Using the same idea with the special conformal generator Z ¼ N dpeÀ

10. Gp ¼ N dd dppdÀ1 eÀ

11. Gp
~
we propose @ @
Vd Z 1 pffiffiffi z
C ¼ t2 pt þ
txi pi þ Mðp2 Þx2 ; ¼ N dppdÀ1 eÀ

12. Gp : (44)
@ 0
pffiffiffiffi
where
, are constants to be determined from the re- Using u ¼

13. Gpz , and the definition of the gamma func-
quirement (38). The solutions are tion, we obtain
1 2
¼À ¼À Vd d
ð1 À Þ
;
ð1 À Þ2
: Z ¼ N pffiffiffiffi d=z À : (45)
z@ ð

14. GÞ z
For the Galilean boost we have In this way, if N is the number of particles, the Helmholtz
Ki ¼ tpi þ Mðp2 Þxi ; free energy is given by
Vd d
and the solution for the constant using the condition (38) F ¼ ÀkTN lnZ ¼ ÀkTN ln N pffiffiffiffi d=z À : (46)
is z@ ð

15. GÞ z
065013-5

16. ROMERO et al. PHYSICAL REVIEW D 81, 065013 (2010)
From this expression we may now obtain all the thermody- Furthermore, it was shown that for our particle model the
namic properties of the system. In particular, the internal ˇ
system has the same local symmetries of Horava gravity.
energy is Another interesting point is that the anisotropic mechanics
@ lnZ d proposed here corresponds to the conformal mechanics of
U ¼ NkT 2 ¼ NkT; (47) [3] for z ¼ 2 and is a generalization of it for arbitrary z.
@T z Also our system includes, in the limit z ! 1, a relativistic
where N is the total number of particles in the gas. particle with Euclidean signature and, in the double scaling
Substituting the number of particles N in terms of the limit z ! 1, m ! 0 with zÀ1 constant, a massless relativ-
m
partition function (Z ¼ N in our normalization) the inter- istic particle. From the equations of motion we showed that
nal energy can be written as there naturally emerges a Finsler type metric that could
Vd ÀðdÞ d ˇ
imply that Horava’s gravity can be related to Finsler ge-
U¼ pffiffiffiffi z ðkTÞðdþzÞ=z : (48) ometry. Also, we studied the dynamical symmetries of our
z@N ð GÞd=z z model and we found all dynamical symmetries of the
Whereas the entropy is ¨
anisotropic Schrodinger equation for arbitrary z that cor-
¨
respond to the generators of the Schrodinger algebra
Z 1 @U Vd ÀðdÞðkÞðdþzÞ=z d d=z d þ z Schrz ðdÞ. We showed that the full Schrodinger algebra
¨
S ¼ dT ¼ pffiffiffiffi
z
ðTÞ :
T @T V z@N ð GÞd=z z d constructed from our generators does not close. However,
(49) we found that a subalgebra SLð2; RÞ indeed closes. An
interesting point is that the explicit realization of the gen-
We conclude from (48) and (49) that the relationship erators is not linear and is analogous to the realization
between energy and entropy in our system is given by given in [6]. As a final point, we remarked that the ther-
STd modynamic properties of our model reproduces the same
U¼ : (50) thermodynamic properties of the recently proposed aniso-
dþz
tropic black branes.
This relationship is exactly that one obtained in the case of
black branes [7]. ACKNOWLEDGMENTS
This work was partially supported under grants
VII. CONCLUSIONS
CONACyT-SEP 55310, CONACyT 50-155I, DGAPA-
In this work we introduced an action invariant under UNAM IN109107, and DGAPA-UNAM IN116408.
anisotropic transformations between spatial and temporal Note added.—While this article was in the peer review-
coordinates. This anisotropic mechanical system is consis- ing process at Physical Review D, several papers addressed
ˇ
tent with the nonrelativistic Horava’s dispersion relation. issues related to our work [26–30].
[1] P. C. Hohenberg and B. I. Halperin, Rev. Mod. Phys. 49, ˇ
[10] P. Horava, Phys. Rev. D 79, 084008 (2009).
435 (1977). [11] P. A. M. Dirac, Lectures on Quantum Mechanics (Dover,
[2] J. L. Cardy, Scaling and Renormalization in Statistical New York, 2001).
Physics (Cambridge University Press, Cambridge, [12] N. Laskin, Phys. Lett. A 268, 298 (2000); Phys. Rev. E 62,
England, 1996). 3135 (2000); 66, 056108 (2002).
[3] V. Alfaro, S. Fubini, and G. Furlan, Nuovo Cimento Soc. [13] P. Horava, Phys. Rev. Lett. 102, 161301 (2009).
Ital. Fis. 34A, 569 (1976). [14] A. M. Mathai, R. K. Saxena, and H. J. Haubold, The H-
[4] R. Britto-Pacumio, J. Michelson, A. Strominger, and A. Function: Theory and Applications (Springer, New York,
Volovich, in Progress in String Theory and M-theory, 2010).
edited by Laurent Baulieu, Michael Green, Marco Picco, [15] M. S. Plyushchay, Mod. Phys. Lett. A 4, 837 (1989); M. S.
and Paul Windey (Kluwer, The Netherlands, 2001), p. 235. Plyushchay and A. V. Razumov, Int. J. Mod. Phys. A 11,
[5] H. E. Camblong and C. R. Ordonez, Phys. Rev. D 68, 1427 (1996).
125013 (2003). [16] M. Henneaux and C. Teitelboim, Quantization of Gauge
[6] M. Henkel, Nucl. Phys. B641, 405 (2002). System (Princeton University Press, Princeton, NJ, 1992).
[7] G. Bertoldi, B. A. Burrington, and A. Peet, Phys. Rev. D [17] C. Teitelboim, Phys. Rev. D 25, 3159 (1982).
80, 126004 (2009). [18] F. Girelli, S. Liberati, and L. Sindoni, Phys. Rev. D 75,
[8] J. L. F. Barbon and C. A. Fuertes, J. High Energy Phys. 09 064015 (2007).
(2008) 030. [19] The Galilei symmetry of the free particle action can be
[9] D. T. Son, Phys. Rev. D 78, 046003 (2008). ¨
central extended to Schrodinger symmetry by the inclu-
065013-6

17. CONFORMAL ANISOTROPIC MECHANICS AND THE . . . PHYSICAL REVIEW D 81, 065013 (2010)
sion of dilatation and a special conformal transformation. [24] C. R. Hagen, Phys. Rev. D 5, 377 (1972); U. Niederer,
See for instance [20]. Helv. Phys. Acta 45, 802 (1972).
[20] C. Leiva and M. S. Plyushchay, Ann. Phys. (N.Y.) 307, 372 [25] K. Balasubramanian and J. McGreevy, Phys. Rev. Lett.
(2003). 101, 061601 (2008).
[21] A. Bagchi and R. Gopakumar, J. High Energy Phys. 07 [26] T. Suyama, arXiv:0909.4833.
(2009) 037. [27] D. Capasso and A. P. Polychronakos, arXiv:0909.5405.
[22] P. D. Alvarez, J. Gomis, K. Kamimura, and M. S. [28] S. K. Rama, arXiv:0910.0411.
Plyushchay, Phys. Lett. B 659, 906 (2008). [29] L. Sindoni, arXiv:0910.1329.
[23] C. Duval, M. Hassaine, and P. A. Horvathy, Ann. Phys. [30] E. Kiritsis and G. Kofinas, J. High Energy Phys. 01 (2010)
(N.Y.) 324, 1158 (2009). 122.
065013-7