Abstract
Consider the shift σ acting on the Bernoulli space \(\varSigma =\{ 1,2,\ldots,n\}^{\mathbb{N}}\). We denote \(\hat{\varSigma }=\{ 1,2,\ldots,n\}^{\mathbb{Z}} =\varSigma \times \varSigma\). We analyze several properties of the maximizing probability μ ∞, A of a Hölder potential \(A:\varSigma \rightarrow \mathbb{R}\). Associated to A(x), via the involution kernel, W(x, y), \(W:\hat{\varSigma }\rightarrow \mathbb{R}\), one can get the dual potential A ∗(y), where \((x,y) \in \hat{\varSigma }\). We denote \(\mu _{\infty,A^{{\ast}}}\) the maximizing probability for A ∗. We would like to consider the transport problem from μ ∞, A to \(\mu _{\infty,A^{{\ast}}}\). In this case, it is natural to consider the cost function c(x, y) = I(x) − W(x, y) +γ, where I is the deviation function for μ ∞, A , as the limit of Gibbs probabilities μ β A for the potential β A when β → ∞. The value γ is a constant which depends on A. We could also take c = −W above. We denote by \(\mathcal{K} = \mathcal{K}(\mu _{\infty,A},\mu _{\infty,A^{{\ast}}})\) the set of probabilities \(\hat{\eta }(x,y)\) on \(\hat{\varSigma }\), such that \(\pi _{x}^{{\ast}}(\hat{\eta }) =\mu _{\infty,A},\,\,\text{and}\,\,\pi _{y}^{{\ast}}(\hat{\eta }) =\mu _{\infty,A^{{\ast}}}\,.\) We describe the minimal solution \(\hat{\mu }\) (which is invariant by the shift on \(\hat{\varSigma }\)) of the Transport Problem, that is, the solution of
The optimal pair of functions for the Kantorovich Transport dual Problem is (−V, −V ∗), where we denote the two calibrated sub-actions by V and V ∗, respectively, for A and A ∗. We show that the involution kernel W is cyclically monotone. In other words, satisfies a twist condition in the support of \(\hat{\mu }\). We analyze the question: is the support of \(\hat{\mu }\) a graph? We also investigate the question of finding an explicit expression for the function \(f:\varSigma \rightarrow \mathbb{R}\) whose c-subderivative determines the graph. We also analyze the same kind of problem for expanding transformations on the circle.
This is a preview of subscription content, log in via an institution.
Buying options
Tax calculation will be finalised at checkout
Purchases are for personal use only
Learn about institutional subscriptionsReferences
Bangert, V.: Mather sets for twist maps and geodesics on tori. In: Dynamics Reported, vol. 1, pp. 1–56. Wiley, Chichester (1988)
Baraviera, A., Lopes, A.O., Thieullen, Ph.: A large deviation principle for equilibrium states of holder potentials: the zero temperature case. Stoch. Dyn. 6, 77–96 (2006)
Baraviera, A.T., Cioletti, L.M., Lopes, A.O., Mohr, J., Souza, R.R.: On the general XY Model: positive and zero temperature, selection and non-selection. Rev. Math. Phys. 23(10), 1063–1113 (2011)
Baraviera, A., Lopes, A.O., Mengue, J.: On the selection of subaction and measure for a subclass of potentials defined by P. Walters. Ergodic Theory Dyn. Syst. 33(5), 1338–1362 (2013)
Baraviera, A., Leplaideur, R., Lopes, A.O.: Ergodic optimization, zero temperature limits and the max-plus algebra. In: Mini-course in XXIX Coloquio Brasileiro de Matemática, Rio de Janeiro (2013)
Bhattacharya, P., Majumdar, M.: Random Dynamical Systems. Cambridge University Press, Cambridge (2007)
Bissacot, R., Garibaldi, E.: Weak KAM methods and ergodic optimal problems for countable Markov shifts. Bull. Braz. Math. Soc. 41(3), 321–338 (2010)
Bousch, T.: Le poisson n’a pas d’arêtes. Ann. Inst. Henri Poincaré Probab. Stat. 36, 489–508 (2000)
Bousch, T.: La condition de Walters. Ann. Sci. l’École Normale Supérieure 34, 287–311 (2001)
Contreras, G.: Ground states are generically a periodic orbit, Arxiv (2013)
Contreras, G., Iturriaga, R.: Global minimizers of autonomous Lagrangians. 22∘ Colóquio Brasileiro de Matemática, IMPA (1999)
Contreras,G., Lopes, A.O., Thieullen, Ph.: Lyapunov minimizing measures for expanding maps of the circle. Ergodic Theory Dyn. Syst. 21, 1379–1409 (2001)
Contreras, G., Lopes, A.O., Oliveira, E.: Ergodic Transport Theory, periodic maximizing probabilities and the twist condition. In: D. Zilberman, A. Pinto (eds.) Modeling, Optimization, Dynamics and Bioeconomy. Springer Proceedings in Mathematics, pp. 183–219. Springer, Cham (2014)
Conze, J.P., Guivarc’h, Y.: Croissance des sommes ergodiques et principe variationnel, manuscript circa (1993)
Delon, J., Salomon, J., Sobolevski, A.: Fast transport optimization for Monge costs on the circle. SIAM J. Appl. Math. 7, 2239–2258 (2010)
Dembo, A., Zeitouni, O.: Large Deviations Techniques and Applications. Springer, New York (1998)
Evans, L., Gomes, D.: Linear programming interpretation of Mather’s variational principle. ESAIM Control Optim. Cal. Var. 8, 693–702 (2002)
Galatolo, S., Pacifico, M.: Lorenz-like flows: exponential decay of correlations for the Poincaré map, logarithm law, quantitative recurrence. Ergodic Theory Dyn. Syst. 30(6), 1703–1737 (2010)
Gangbo, W., McCann, R.J.: The geometry of optimal transportation. Acta Math. 177, 113–161 (1996)
Garibaldi, E., Lopes, A.O.: Functions for relative maximization. Dyn. Syst. 22, 511–528 (2007)
Garibaldi, E., Lopes, A.O., Thieullen, Ph.: On calibrated and separating sub-actions. Bull. Braz. Math. Soc. 40(4), 577–602 (2009)
Garibaldi, E., Lopes, A.O.: The effective potential and transshipment in thermodynamic formalism at temperature zero. Stoch. Dyn. 13(1), 1250009 (13 p) (2013)
Garibaldi, E., Thieullen, Ph.: Minimizing orbits in the discrete Aubry-Mather model. Nonlinearity 24(2), 563–611 (2011)
Garibaldi, E., Thieullen, Ph.: Description of some ground states by Puiseux technics. J. Stat. 146(1), 125–180 (2012)
Gole, C.: Sympletic Super-Twist Maps. World Scientific, Singapore (1998)
Hunt, B.R., Yuan, G.C.: Optimal orbits of hyperbolic systems. Nonlinearity 12, 1207–1224 (1999)
Jenkinson, O.: Ergodic optimization. Discrete Continuous Dyn. Syst. Ser. A 15, 197–224 (2006)
Jenkinson, O.: Every ergodic measure is uniquely maximizing. Discrete Continuous Dyn. Syst. Ser. A 16, 383–392 (2006)
Jenkinson, O.: A partial order on x2 -invariant measures. Math. Res. Lett. 15(5), 893–900 (2008)
Jenkinson, O.: Optimization and majorization of invariant measures. Electron. Res. Announc. Am. Math. Soc. 13, 1–12 (2007)
Jenkinson, O., Steel, J.: Majorization of invariant measures for orientation-reversing maps. Ergodic Theory Dyn. Syst. 30(5), 1471–1483 (2010)
Kloeckner, B.: Optimal Transport and dynamics of circle expanding maps acting on measures. Ergodic Theory Dyn. Syst. 33(2), 529–548 (2013)
Kloeckner, B., Lopes, A.O., Stadlbauer, M.: Contraction in the Wasserstein metric for some Markov chains, and applications to the dynamics of expanding maps, preprint (2014)
Kloeckner, B., Giulietti, P., Lopes, A.O., Marcon, D.: On the Geometry of Thermodynamical Formalism, preprint (2014)
Leplaideur, R.: A dynamical proof for the convergence of Gibbs measures at temperature zero. Nonlinearity 18(6), 2847–2880 (2005)
Lopes, A., Mengue, J.: Duality theorems in ergodic transport. J. Stat. Phys. 149(5), 921–942 (2012)
Lopes, A.O., Oliveira, E.R.: On the thin boundary of the fat attractor, preprint UFRGS (2011)
Lopes, A.O., Thieullen, Ph.: Sub-actions for Anosov diffeomorphisms. Astérisque 287, 135–146 (2003)
Lopes, A.O., Thieullen, P.: Mather measures and the Bowen-Series transformation. Ann. Inst. Henri Poincaré Analyse non Linéaire 23, 663–682 (2006)
Lopes, A.O., Oliveira, E.R., Smania, D.: Ergodic transport theory and piecewise analytic subactions for analytic dynamics. Bull. Braz. Math. Soc. 43(3), 467–512 (2012)
Lopes, A.O., Mengue, J.K., Mohr, J., Souza, R.R.: Entropy, pressure and duality for Gibbs plans in ergodic transport. Bull. Braz. Math. Soc. (to appear)
Lopes, A., Mengue, J.K., Mohr, J., Souza, R.R.: Entropy and variational principle for one-dimensional lattice systems with a general a-priori probability: positive and zero temperature. Ergodic Theory Dyn. Syst. (to appear)
Mather, J.: Action minimizing invariant measures for positive definite Lagrangian Systems. Math. Z. 207(2), 169–207 (1991)
Mengue, J.K., Oliveira, E.R.: Duality results for Iterated Function Systems with a general family of branches, preprint Arxiv (2014)
Mitra, T.: Introduction to dynamic optimization theory. In: Majumdar, M., Mitra, T., Nishimura, K. (eds.) Optimization and Chaos. Studies in Economic Theory. Springer, Heidelberg (2000)
Morris, I.D.: A sufficient condition for the subordination principle in ergodic optimization. Bull. Lond. Math. Soc. 39(2), 214–220 (2007)
Parry, W., Pollicott, M.: Zeta functions and the periodic orbit structure of hyperbolic dynamics. In: Astérisque vol. 187–188. Société mathématique de France, Paris (1990)
Rachev, S., Ruschendorf, L.: Mass Transportation Problems, vol. 1 and 2. Springer, New York (1998)
Ruscheendorf, L.: On c-optimal random variables. Stat. Probab. Lett. 27, 267–270 (1996)
Savchenko, S.V.: Cohomological inequalities for finite Markov chains. Funct. Anal. Appl. 33, 236–238 (1999)
Souza, R.R.: Ergodic and Thermodynamic Games, preprint (to appear in Stochastics and Dynamics, 2014)
Tal, F.A., Zanata, S.A.: Maximizing measures for endomorphisms of the circle. Nonlinearity 21, 2347–2359 (2008)
Villani, C.: Topics in Optimal Transportation. AMS, Providence (2003)
Villani, C.: Optimal Transport: Old and New. Springer, Berlin (2009)
Acknowledgements
The first author is partially supported by CNPq, CAPES and FAPERGS.
Author information
Authors and Affiliations
Corresponding author
Editor information
Editors and Affiliations
Appendix
Appendix
Here we consider first the shift \(\varSigma =\{ 0,1\}^{\mathbb{N}}\), and Σ as a metric space with the usual distance:
Additionally, we suppose that Σ is ordered by x < y, if x i = y i for i = 1. . n − 1, and x n = 0 and y n = 1.
As the usual, we consider the dynamical system (Σ, σ) where σ: Σ → Σ is given by σ(x) = σ(x 1, x 2, x 3, …) = (x 2, x 3, x 4, …).
-
(a)
Potentials and the Involution Kernel
As usual we denote
$$\displaystyle{\tau _{x}^{{\ast}}(y) = (x_{ 1},y_{1},y_{2},y_{3},\ldots )\text{ and }\tau _{y}(x) = (y_{1},x_{1},x_{2},x_{3},\ldots ),}$$and
$$\displaystyle{\hat{\sigma }(x,y) = (\sigma (x),\tau _{x}^{{\ast}}(y))\text{ and }\hat{\sigma }^{-1}(x,y) = (\tau _{ y}x,\sigma ^{{\ast}}(y)),}$$the skew product map, where σ ∗(y = (y 1, y 2, y 3, …)) = (y 2, y 3, y 4, …).
We also define τ k, y x = (y k , y k−1, … y 2, y 1, x 0, x 1, x 2, …), where x = (x 0, x 1, x 2, …), y = (y 1, y 2, y 3, …). In a similar way we define τ k, y ∗ x.
Given a continuous function \(A:\varSigma \rightarrow \mathbb{R}\), remember that a continuous function \(W:\varSigma \times \varSigma \rightarrow \mathbb{R}\) is an involution kernel for A if \((W \circ \hat{\sigma }^{-1} - W + A \circ \hat{\sigma }^{-1})(x,y)\) does not depends on x; In this case the continuous function \(A^{{\ast}}(y) = (W \circ \hat{\sigma }^{-1} - W + A \circ \hat{\sigma }^{-1})(x,y)\) is called the W-dual potential of A.
As in [2] we define the cocycle Δ A (x, x′, y), where
and its dual version \(\varDelta _{A^{{\ast}}}(x,y,y')\), where
Note that:
-
(i)
Δ A (x, x′, y) = −Δ A (x′, x, y), in particular Δ A (x, x, y) = 0,
-
(ii)
Δ A (x, x′, y) +Δ A (x′, x″, y) = Δ A (x, x″, y),
-
(iii)
Δ A (x, x′, y) = Δ A (τ y x, τ y x′, σ ∗(y)) + [A ∘τ y x − A ∘τ y x′],
and the same relations are true for \(\varDelta _{A^{{\ast}}}(x,y,y')\).
Using this properties one can prove that, for any involution kernel we have \(W(x,y) - W(x',y) =\varDelta _{A}(x,x',y)\text{ and }W(x,y) - W(x,y') =\varDelta _{A^{{\ast}}}(x,y,y').\)
From this fact, we get that the difference between two involution kernels for A is a continuous function of y: {Involution kernels for A}∕C0(Σ) = W 0, where W 0(x, y) = Δ A (x, x′, y) for a fix x′ ∈ Σ is called a fundamental involution kernel of A. Indeed, the property (iii) shows that W 0 is an involution kernel for A.
On the other hand, given another involution kernel, W we have W(x, y) − W(x′, y) = Δ A (x, x′, y), thus
where g(y) = W(x′, y) ∈ C 0(Σ).
As an example we compute the general dual potential. First for W 0(x, y) = Δ A (x, x′, y) we get:
Given another involution kernel, W we have W(x, y) = W(x′, y) + W 0(x, y) thus
-
(b)
The Twist Property of an Involution Kernel
If \(A:\varSigma \rightarrow \mathbb{R}\) is a potential and W an arbitrary involution kernel for A, as we said before, W has the twist property, if for any, a, b, a′, b′ ∈ Σ
provided that a < a′ and b < b′.
If we rewrite this inequality as,
we get an alternative criteria for the twist property, that is, W has the twist property, if for any, a, a′ ∈ Σ the function y → Δ A (a, a′, y), is strictly increasing, provided that a < a′.
Remark 5
This characterization shows a very important fact. The twist property is a property of A, so we can said that A is a twist potential or equivalently A has a twist involution kernel (as, obviously other involution kernel is also twist).
Remark 6
As an initial approximation we can consider a different setting of dynamics. Let T(x) = −2x ( mod 1 ), and
the inverse branches that defines the skew maps (that are not the actual natural extension of T):
So, one can compute an involutive (that is, A ∗(y) = A(y)) smooth kernel for A 1(x) = x and A 2(x) = x 2 given by
As a corollary we get that any potential A(x) = a + bx + cx 2 has a smooth involution kernel given by W(x, y) = a + bW 1(x, y) + cW 2(x, y).
Here and in the next paragraphs, we will denote
where A(x) = a + bx + cx 2 is a polynomial of degree 2.
We observe that the twist property can be derived from the positivity of the second mix derivative of the involution kernel when it is smooth. Note that,
thus W 1 is not twist and W 2 is. Actually any potential A(x) = a + bx + cx 2 where c > 0 is twist.
Remark 7
In this remark we are going to consider the case of A(x) = a + bx + cx 2 where c < 0 (not twist). In this case we will be able to compute the calibrated subaction explicitly, which, we believe, it is interesting in itself.
As a first example consider A(x) = −(x − 1)2 which is a convex potential.
From [30, 31] we get that the unique maximizing measure for this potential is μ ∞ = δ 2∕3, so the critical value is m = A(2∕3). Using the fact that m = A(2∕3) one can show that there is a unique (up to constants) calibrated subaction ϕ given by:
where the kernel is given by
As a second example consider \(A(x) = -(x -\frac{1} {2})^{2}\) which it is also a concave potential.
The general arguments in [31] shown that any maximizing measure for this potential is μ ∞ = (1 − t)δ 1∕3 + t δ 2∕3, where t ∈ [0, 1], so the critical value is m = A(1∕3) = A(2∕3). In this case the involutive smooth involution kernel is:
It is easy to verify that,
is indeed a calibrated subaction for A, where
V 1(x) = W(x, 1∕3) − W(1∕3, 1∕3) = Δ(x, 1∕3, 1∕3) = −(1∕3)x 2 + (1∕9)x,
V 2(x) = W(x, 2∕3)−W(2∕3, 2∕3) = Δ(x, 2∕3, 2∕3) = −(1∕3)x 2+(5∕9)x−2∕9,
Note that,
Thus ϕ(τ 0 x) + A(τ 0 x) − m = V 1(x). Analogously, ϕ(τ 1 x) + A(τ 1 x) − m = V 2(x) so
-
(c)
Twist Criteria
Is natural to consider a criteria for the twist property for a class of functions that has a small dependence on the cubic (or higher order) terms. Let P 2 + = { p(x) = a + bx + cx 2 | c > 0} be the set of strictly convex polynomial. Consider p ∈ P 2 +, and define
Theorem 8
For any p ∈ P 2 + , there exists \(\varepsilon > 0\) such that all \(A \in \mathcal{C}_{\varepsilon }(p)\) is twist.
Proof
Consider p ∈ P 2 + fixed. So, p has a smooth and involutive involution kernel given by
that is, p ∗(y) = p(y), where \(W_{1}(x,y) = -\frac{1} {3}(x + y)\) and \(W_{2}(x,y) = \frac{1} {3}(x^{2} + y^{2}) -\frac{4} {3}xy\), are the involution kernel associated to x and x 2 respectively. Let, \(A = p +\varepsilon R \in \mathcal{C}_{\varepsilon }(p)\), and W R be the involution kernel for R. Since R is C3 we get that, its corresponding involution kernel W R is C2 in the variable x. Using the linearity of the cohomological equation, we get \(W_{A}(x,y) = p(W)(x,y) +\varepsilon W_{R}(x,y)\), and differentiating with respect to x, we have
Since \(-\frac{4} {3}c < 0\), and \(\frac{\partial } {\partial x}W_{R}(x,y) \in \text{C}^{0}([0,1]^{2})\) the compactness of [0, 1]2 implies that \(\frac{\partial } {\partial x}W_{A}(x,\cdot )\) is a strictly decreasing function for any \(\varepsilon\) small enough, which is sufficient to ensure the twist property.
Remark 8
If, A ∈ C∞([0, 1]) is strongly convex, we can consider a perturbation of A of order 2 given by
where \(p_{A} = A(0) - A'(0)x + \frac{A''(0)} {2} x^{2} \in P_{ 2}^{+}\). Thus, we can find \(\varepsilon _{0} > 0\) such that \(B_{\varepsilon }\) is twist for any \(0 <\varepsilon <\varepsilon _{0}\).
-
(d)
The Involution Kernel is Bi-Hölder
We consider now T(x) = 2x (mod 1) on the interval [0, 1] and the shift σ on \(\varOmega =\{ 0,1\}^{\mathbb{N}}\). A natural question is the regularity of the involution kernel W. We denote τ j , j = 0, 1 the two inverse branches of T. Given \(w = (w_{1},w_{2},\ldots ) \in \{ 0,1\}^{\mathbb{N}}\) we denote by τ k, w the transformation in [0, 1] given by \(\tau _{k,w}(x) = (\tau _{w_{k}} \circ \tau _{w_{k-1}} \circ \,\ldots \,\circ \tau _{w_{1}})\,(x).\) We have that, for a fixed x 0
and, the involution kernel W can be described as: for any (x, w) we have W(x, w) = Δ(x, x 0, w). It is easy to see that W is Hölder on the variable x. Consider a, b ∈ Ω and suppose that d(a, b) = 2−n. In this way a j = b j , j = 1, 2…, n − 1, n. We denote \(\bar{a} =\sigma ^{n}(a)\) and \(\bar{b} =\sigma ^{n}(b)\).
Proposition 7
Suppose A is α−Hölder. Consider a,b ∈Ω such that d(a,b) = 2 −n . For a fixed x ∈ [0,1] we have | W(x,a) − W(x,b) |≤ C (2 −n ) α .
Proof
Note that for z = τ n, a (x) = τ n, b (x) and z 0 = τ n, a (x 0) = τ n, b (x 0) we have
Note also that | z − z 0 | ≤ d(a, b) = 2−n. Consider z = z 0 + h, then
Then,
From the above we get:
Theorem 9
If \(A: S^{1} \rightarrow \mathbb{R}\) is Hölder then \(W: S^{1} \times \{ 0,1\}^{\mathbb{N}} \rightarrow \mathbb{R}\) is bi-Hölder.
-
(e)
The Fenchel-Rockafellar Theorem Given \(f: \mathbb{R} \rightarrow \mathbb{R}\) defined on the variable x, the Legendre transform of f, denoted by f ∗, is the function on the variable p defined by
$$\displaystyle{f^{{\ast}}(p) =\sup _{ x\in \mathbb{R}}\{p\,x -\, f(x)\}.}$$
Theorem 10 (Fenchel-Rockafellar)
Suppose f(x) is smooth strictly convex, \(f: \mathbb{R} \rightarrow \mathbb{R}\) , and, g(x) is smooth strictly concave, \(g: \mathbb{R} \rightarrow \mathbb{R}\) . Denote by f ∗ and g ∗ the corresponding Legendre transforms on the variable p. Then,
Proof
By convexity and concavity properties we have that there exists x 0 such that
It is also true that f′(x 0) − g′(x 0) = 0. Denote by \(\overline{p}\) that value \(\overline{p} = f'(x_{0})\). We illustrate the proof via two pictures in a certain particular case. Figure 11 shows a geometric picture of the position and values of f(x 0) − g(x 0), \(g^{{\ast}}(\overline{p})\) and \(f^{{\ast}}(\overline{p}).\) Note that in this picture we have that f(x 0) − g(x 0) > 0. This picture also shows the graph of \(\overline{p}\,x\) as a function of x. We observe that the Legendre transform is not linear on the function. Let’s consider different values of p and estimate f ∗(p) and g ∗(p). Suppose first \(p\, >\, \overline{p}\). In Fig. 12 we show the graph of p x, and the values of f ∗(p) and g ∗(p). We denote by x 2 the value such that
Note that x 2 > x 0. We denote by x 1 the value such that
Note that x 1 < x 0.
Note also that f ∗(p) and g ∗(p) have different signs. From this picture one can see that g ∗(p) − f ∗(p) < f(x 0) − g(x 0). In the case \(p\, <\, \overline{p}\) a similar reasoning can be done.
Rights and permissions
Copyright information
© 2015 Springer International Publishing Switzerland
About this paper
Cite this paper
Lopes, A.O., Oliveira, E.R., Thieullen, P. (2015). The Dual Potential, the Involution Kernel and Transport in Ergodic Optimization. In: Bourguignon, JP., Jeltsch, R., Pinto, A., Viana, M. (eds) Dynamics, Games and Science. CIM Series in Mathematical Sciences, vol 1. Springer, Cham. https://doi.org/10.1007/978-3-319-16118-1_20
Download citation
DOI: https://doi.org/10.1007/978-3-319-16118-1_20
Publisher Name: Springer, Cham
Print ISBN: 978-3-319-16117-4
Online ISBN: 978-3-319-16118-1
eBook Packages: Mathematics and StatisticsMathematics and Statistics (R0)