Abstract
We consider an open interacting particle system on a finite lattice. The particles perform asymmetric simple exclusion and are randomly created or destroyed at all sites, with rates that grow rapidly near the boundaries. We study the hydrodynamic limit for the particle density at the hyperbolic space-time scale and obtain the entropy solution to a boundary-driven quasilinear conservation law with a source term. Different from the usual boundary conditions introduced in Bardos et al (Commun Partial Differ Equ 4(9):1017–1034, https://doi.org/10.1080/03605307908820117, 1979) and Otto (C R Acad Sci Paris 322(1):729–734, 1996), discontinuity (boundary layer) does not formulate at the boundaries due to the strong relaxation scheme.
Similar content being viewed by others
Avoid common mistakes on your manuscript.
1 Introduction
In the past decades, hydrodynamic limit for interacting particle system with boundary effect has attracted a lot of attention [1, 2, 4,5,6, 8, 11, 15, 24]. The limit captures the evolution of the conserved field of the microscopic dynamics as hydrodynamic equation, at the macroscopic time scale that the dynamics is equilibrated locally. Most of these works focus on symmetric dynamics and the Fick’s law of diffusion or fractional diffusion with various types of boundary conditions, see, e.g., [2, 4,5,6, 8, 15].
For asymmetric mass-conserving systems, the dynamics reaches local equilibrium at the hyperbolic time scale, and the hydrodynamic equations are given by hyperbolic transport equations [22]. When nonlinear interaction exists, these equations are featured by discontinuous phenomenon both inside the domain (shock wave) and at the boundary (boundary layer). The non-regularity becomes the main obstacle in deducing the hydrodynamic limit. Asymmetric simple exclusion process (ASEP) with open boundaries is the simplest model. In its dynamics, each particle performs an asymmetric random walk on the finite lattice \(\{1,\ldots ,N-1\}\) under the exclusion rule: two particles cannot occupy the same site simultaneously. Particles are created and annihilated randomly at sites 1 and \(N-1\), modeling the exchange of mass between the system and two external reservoirs at given densities. In [1, 24, 25], the hydrodynamic limit for the mass density of open ASEP is proved to be Burgers equation with boundary conditions introduced in [3, 20]. Due to the discontinuous nature, these boundary conditions do not prescribe the density at boundary, even when the reservoir dynamics is overwhelmingly accelerated compared to the exclusion [25]. Instead, they impose a set of possible values for the boundary density. The hydrostatic limit for the same dynamics is studied in [1, 9]: the stationary density profile is the stationary solution to the hydrodynamic equation. It is determined by the boundary data through a variational property [21]: the stationary flux is maximized if the density gradient is opposite to the drift, and is minimized otherwise. The result is generalized in [7] to the quasi-static transform: if the reservoir densities are changing slowly at a time scale that is larger than the hyperbolic one, the profile evolves with the corresponding quasi-static solution [17].
The motivation of this article is to study the hydrodynamic limit for a hyperbolic system disturbed by a nonlocal external field. We are particularly interested in the macroscopic behavior when the perturbation is extremely strong at the boundary. Consider the ASEP on \(\{1,\ldots ,N-1\}\) where particles are created (resp. annihilated) at each site i with rate \(V_i\rho _i\) (resp. \(V_i(1-\rho _i)\)). Assume two profiles \((V,\rho ):(0,1)\rightarrow {\mathbb {R}}_+\times [0,1]\) such that
In other words, a reservoir of density \(\rho _i\) is placed at each site i, and the system exchanges particles with it with frequency \(V_i\) that is growing rapidly near the boundaries. When the exclusion dynamics is accelerated by N, the density profile shall evolve with the \(L^\infty \) entropy solution to the following quasilinear balance law in the [0, 1]-interval:
with proper boundary conditions. We prove in Theorem 2.8 that, when the integrals of V are infinity around both 0 and 1, the boundary conditions are \(u|_{x=0}=\rho (0)\), \(u|_{x=1}=\rho (1)\). In sharp contrast to the equations obtained in [1, 24, 25], the boundary values of u are fixed by \(\rho \) in a weak sense, see Proposition 2.11. Hence, any shock wave is attenuated while approaching the boundaries, and no boundary layer is observable at any positive macroscopic time. A consequence of the hydrodynamic limit is the \(L^1\)-weak continuity in time of the entropy solution obtained in Corollary 2.10.
The term G in (1.2) acts as a source (resp. sink) where u is less (resp. greater) than \(\rho \), so it can be viewed as a relaxation scheme to the profile \(\rho \). When \(\rho \) is a constant, it is a conservation system with relaxation introduced in [16], with the first component degenerated to a stationary solution. Such system is widely used to model non-equilibrium transport in kinetic theory and fluid dynamics. In our situation, the entropy solution to the initial-boundary problem of (1.2) is constructed in different ways depending on the integrability of V, see Definition 2.1 and 2.7. We focus on the non-integrable case and discuss the integrable case briefly in Sect. 2.4.
The proof in this article is proceeded in two main steps. First, we prove that in the space-time scaling limit, the empirical Young measure of the particle field is concentrated on the space of Dirac-type Young measures. Then, we show that the limit is a measure-valued entropy solution to (1.2) with proper boundary conditions. The hydrodynamic limit then follows from the uniqueness of the entropy solution. Both steps are proved through delicate analyses of the microscopic entropy production associated with Lax entropy–flux pairs.
The use of Young measure and microscopic entropy production is present in the seminal paper [22]. It is combined with the compensated compactness method to prove the concentration property of the Young measure in [12, 13]. To use this method, additional oscillating dynamics is added to ASEP to create microscopic viscosity. Finally, we point out that although the process studied in this article is attractive, we cannot apply the coupling argument used in [1] because the invariant measure is not product in general.
2 Model and Results
2.1 Model
For a scaling parameter \(N \in {\mathbb {N}}_+\), consider the configuration space
The dynamics on \(\Omega _N\) consists of three parts: the nearest-neighbor asymmetric exclusion, the external Glauber field and the boundary dynamics. The exclusion is generated by
for any function f on \(\Omega _N\), where, for constant \(p\in (\tfrac{1}{2},1]\),
\(\sigma _N\) is a parameter that grows to infinity slower than N, and \(\eta ^{i,i+1}\) is the configuration obtained from \(\eta \) by swap** the values of \(\eta _i\) and \(\eta _{i+1}\). The factor \(\sigma _N\) stands for a strong microscopic viscosity, which is necessary for the technique used in Sect. 5. The Glauber dynamics is generated by
where, for parameters \(V_i>0\) and \(\rho _i\in (0,1)\),
and \(\eta ^{i}\) is the configuration obtained from \(\eta \) by flip** the value of \(\eta _i\). Finally, the sites \(i=0\) and N are attached to two extra birth-and-death dynamics, interpreted as boundary reservoirs. The corresponding generator reads
where, for boundary rates \(c_{\textrm{in}}^\pm \), \(c_{\textrm{out}}^\pm \ge 0\),
Assume two profiles \(V:(0,1)\rightarrow {\mathbb {R}}_+\) and \(\rho :[0,1]\rightarrow (0,1)\) such that \(V_i=V(\tfrac{i}{N})\) and \(\rho _i=\rho (\tfrac{i}{N})\) for \(i=1\),..., \(N-1\). Suppose that \(V\in {\mathcal {C}}^1((0,1);{\mathbb {R}}_+)\), \(V\rightarrow +\infty \) as \(x\rightarrow 0\), 1, and \(\rho \in {\mathcal {C}}^1([0,1];(0,1))\) with Lipschitz continuous \(\rho '\). In particular,
The generator of the process then reads
where the factor N corresponds to the hyperbolic time scale.
2.2 Scalar Balance Law in a Bounded Domain
In this part, we introduce the partial differential equation that is obtained in the hydrodynamic limit for the model defined in the previous section. Let
be the macroscopic flux and the source term corresponding to \(L_\textrm{exc}\) and \(L_{\textrm{G}}\), respectively. Given measurable function \(u_0:(0,1)\rightarrow [0,1]\), consider the following balance law: for \((t,x)\in \Sigma :={\mathbb {R}}_+\times (0,1)\),
with proper boundary conditions that will be specified later.
The weak solution to (2.7) is in general not unique, so we are forced to consider the entropy solution. Recalling (1.1), our aim is to examine the case when the strength of the source is extremely strong at the boundaries. We see in Definitions 2.1 and 2.7 that the definition of entropy solution is different when V is integrable or not at the boundaries.
We begin with the case that V is non-integrable at 0 and 1, i.e., for any small y,
Recall that a Lax entropy–flux pair of (2.7) is a pair of functions f, \(q \in \mathcal C^2({\mathbb {R}})\) such that \(f''\ge 0\) and \(q'=J'f'=(2p-1)(1-2u)f'(u)\) for all \(u\in {\mathbb {R}}\).
Definition 2.1
Suppose that V satisfies (2.8). We call \(u=u(t,x)\) an entropy solution to (2.7) with the compatible boundary conditions
if \(u:\Sigma \rightarrow [0,1]\) is measurable and satisfies the generalized entropy inequality
for any Lax entropy–flux pair (f, q) and any \(\varphi \in {\mathcal {C}}_c^2 ({\mathbb {R}}\times (0,1))\), \(\varphi \ge 0\).
Remark 2.2
When \(\rho \in C^1\), u in Definition 2.1 satisfies the energy estimate
Indeed, suppose that \(\rho \) is smooth. For any \(\varepsilon >0\), choose \(\psi _\varepsilon \in {\mathcal {C}}_c^\infty ((0,1))\) such that \(\psi _\varepsilon (x)\in [0,1]\), \(\psi _\varepsilon |_{[\varepsilon ,1-\varepsilon ]}\equiv 1\) and \(|\psi '_\varepsilon (x)|\le 2\varepsilon ^{-1}\). Fixing any \(\phi \in {\mathcal {C}}_c^\infty ({\mathbb {R}})\) such that \(\phi \ge 0\) and applying (2.10) with \(f_1=\tfrac{1}{2}u^2\), \(\varphi _1=\phi (t)\psi _\varepsilon (x)\) and \(f_2=-u\), \(\varphi _2=\varphi _1\rho \) respectively, we obtain the upper bound
where \(q_1\) is the flux corresponding to \(f_1\). Since \(|\psi _\varepsilon |\le 1\), the first term on the right-hand side is bounded by \(|\phi |_\infty \Vert f_1(u_0)-u_0\rho \Vert _{L^\infty }\). The second term reads
where \(C=\Vert f_1(u)-u\rho \Vert _{L^\infty } + \Vert q_1(u)-J(u)\rho \Vert _{L^\infty } + \Vert J(u)\Vert _{L^\infty }\). Since \(|\psi _\varepsilon |\le 1\), \(|\psi '_\varepsilon |\le 2\varepsilon ^{-1}\) and is non-zero if and only if \(x\in (0,\epsilon )\cup (1-\epsilon ,1)\), it is bounded by \(C_\phi (1+|\rho '|_\infty )\) with a constant \(C_\phi \) that is independent of \(\varepsilon \). Taking \(\varepsilon \rightarrow 0\) and using monotone convergence theorem,
Since \(\phi \in {\mathcal {C}}_c^\infty ({\mathbb {R}};{\mathbb {R}}_+)\) is arbitrary, (2.11) holds for any finite \(T>0\). By standard argument of compactness, the estimate can be extended to any \(\rho \in C^1([0,1])\).
Remark 2.3
If u is continuous in space, (2.11) together with (2.8) implies that \(u(t,0)=\rho (0)\) and \(u(t,1)=\rho (1)\) for almost all \(t>0\). Hence, (2.9) turns out to be the reasonable choice of the boundary conditions, see also Proposition 2.11 below.
The following uniqueness criteria is taken from [26, Theorem 2.12].
Proposition 2.4
Assume further that
Then, there is at most one function \(u \in L^\infty (\Sigma )\) that fulfills Definition 2.1.
Remark 2.5
Suppose that \(V>0\) satisfies (2.12). By Cauchy–Schwarz inequality,
which means that V is not integrable at 0. The same argument holds for the integration on \((1-y,1)\). Therefore, (2.12) contains the non-integrable condition (2.8).
Now we turn to the integrable case: \(V \in L^1((0,1))\). The next definition is first introduced by F. Otto, see [20, Eq. 9].
Definition 2.6
We call \((F, Q) \in {\mathcal {C}}^{2}([0,1]^2; {\mathbb {R}}^{2})\) a boundary entropy–flux pair if
-
(i)
for all \(k \in [0,1]\), \((F,Q) (\cdot , k)\) is a Lax entropy-flux pair, i.e., \(\partial _uQ(\cdot ,k)=J'\partial _uF(\cdot ,k)\);
-
(ii)
for all \(k \in [0,1]\), \(F(k, k)=\partial _{u} F(u, k)|_{u=k}=Q(k, k)=0\).
For V integrable, the definition of entropy solution is the same as [20, Proposition 2] for \(V\equiv 0\) and [18, Definition 1] for V bounded and smooth.
Definition 2.7
Let \(\alpha \), \(\beta \in [0,1]\) be two constants and suppose that \(V \in L^1 ((0,1))\). We call \(u=u(t,x)\) an entropy solution to (2.7) with the boundary conditions given by
if \(u: \Sigma \rightarrow [0,1]\) is a measurable function such that for any boundary entropy–flux pair (F, Q), any \(k \in [0,1]\), and any \(\varphi \in {\mathcal {C}}_c^2 ({\mathbb {R}}^2)\) such that \(\varphi \ge 0\),
where \((f_k,q_k):=(F,Q)(\cdot ,k)\).
Since the integrable case is not the focus of this paper, we omit the uniqueness and other properties and refer to [26] and the references therein.
2.3 Hydrodynamic Limit
Let \(\{\eta ^N(t)\in \Omega _N;t\ge 0\}\) be the Markov process generated by \(L_N\) in (2.5) and initial distribution \(\mu _N\). Through this article, the superscript N in \(\eta ^N\) is omitted when there is no confusion. Denote by \({\mathbb {P}}_{\mu _N}\) the distribution of \(\eta (\cdot )\) on \({\mathcal {D}}([0,\infty ),\Omega _N)\), the space of all càdlàg paths on \(\Omega _N\), and by \({\mathbb {E}}_{\mu _N}\) the expectation of \({\mathbb {P}}_{\mu _N}\).
Suppose that the sequence of \(\mu _N\) is associated with a measurable function \(u_0:(0,1)\rightarrow [0,1]\) in the following sense: for any \(\psi \in {\mathcal {C}}({\mathbb {R}})\),
Our main result shows that in the non-integrable case, the empirical density of the particles converges, as \(N\rightarrow \infty \), to the entropy solution to (2.7) and (2.9).
Theorem 2.8
Assume (2.12), (2.13) and (2.16). Also assume that
Then, for any \(\psi \in {\mathcal {C}}({\mathbb {R}})\) and almost all \(t > 0\),
where u is the unique entropy solution in Definition 2.1.
Remark 2.9
Below we list two important remarks concerning Theorem 2.8.
-
(i)
We assume (2.12) and (2.13) only for the uniqueness in Proposition 2.4. If V satisfies only (2.8), our argument proves that the empirical distribution of \(\eta ^N\) is tight and all limit points are concentrated on the possible entropy solutions.
-
(ii)
Observe that the rates in \(L_{\textrm{bd}}\) do not appear in the limit. Indeed, let \(F_\epsilon (\eta ) = \sum _{0 \le i \le \epsilon N} \eta _i\) be the cumulative mass on \(\{\eta _0,\ldots ,\eta _{[\epsilon N]}\}\). Then,
$$\begin{aligned} L_N F_\epsilon (\eta ) =\,&N\big [c_{\textrm{in}}^- - (c_{\textrm{in}}^-+c_{\textrm{out}}^-)\eta _0\big ] -NJ_{[\epsilon N]}\\&- N(1-p+\sigma _N)(\eta _{[\epsilon N]}-\eta _{[\epsilon N]+1}) + \sum _{1 \le i \le \epsilon N} V_i(\rho _i-\eta _i), \end{aligned}$$where \(J_i=(2p-1)\eta _i(1-\eta _{i+1})\). From (2.8), \(\sum _{1 \le i \le \epsilon N} V_i \gg N\). Hence, to make the contribution of the last term be of order \({\mathcal {O}}(N)\), the mass density of the boundary block \(\{1,\ldots ,\epsilon N\}\) should be near to \(\rho (0)\).
As a corollary of Theorem 2.8, the regularity of the entropy solution is improved.
Corollary 2.10
Assume (2.12) and (2.13). Let u be the unique entropy solution to (2.7) and (2.9) in Definition 2.1. Then,
where \(L^1=L^1((0,1))\) is endowed with the weak topology. In particular, the convergence in Theorem 2.8 holds for all \(t>0\).
Under (2.12), the macroscopic density near the boundary is prescribed by the reservoir in the following sense: for any \(t>0\),
and similarly for the right boundary. Indeed, for any \(t>0\), \(y\in (0,1)\) and \(\delta >0\),
Taking \(y\rightarrow 0+\), (2.11) together with (2.12) yields that
As \(\delta \) is arbitrary, the limit is 0. Recall that \(\rho \) is continuous, so we have
Combining this with Theorem 2.8, we obtain (2.19) for all positive time t. These limits can be derived directly from the microscopic dynamics by imposing a slightly stronger growth condition on V, see the next proposition.
Proposition 2.11
Suppose that V satisfies the following condition:
Then, (2.19) and the similar limit for the right boundary hold for all \(t > 0\).
Example
Fix some \(\gamma >0\), \(\rho _0\) and \(\rho _1\in (0,1)\). By taking
we obtain the source term given by
In this case, the dynamics of \(L_{\textrm{G}}\) can be interpreted as two infinitely extended reservoirs [4,5,6] placed respectively at the sites \(\{-1,-2,\ldots ,\}\) and \(\{N+1,N+2,\ldots \}\). When \(\gamma \ge 1\), V satisfies (2.12), so the hydrodynamic limit can apply.
2.4 Discussion on the Integrable Case
When \(V \in L^1 ((0,1))\), we expect that Theorem 2.8 holds with the entropy solution in Definition 2.7. Since the dynamics of \(L_{\textrm{G}}\) is no more dominating at the boundaries, the boundary data \((\alpha ,\beta )\) may depend on \(c_{\textrm{in}}^\pm \), \(c_{\textrm{out}}^\pm \) as well as V, \(\rho \). In particular when \(V=0\), we expect that \(\alpha \), \(\beta \) are determined by
This is proved in [1] for microscopic dynamics without extra symmetric regularization and the special choice of reservoirs such that
We underline that the problem remains open for general reservoirs even when \(V=0\).
The situation is easier when further speed-up is imposed on the boundary reservoirs. Let \(V \in L^1((0,1))\) satisfy (1.1) and assume the compatibility conditions
Fix \(a>0\) and consider the process generated by \(L'_N:= N(L_{\textrm{exc}} + L_{\textrm{G}} + N^aL_{\textrm{bd}})\). In this case, the hydrodynamic equation is still given by (2.7) and (2.9), but the solution should be understood in the sense of Definition 2.7. This can be proved with the argument in [25].
3 Outline of the Proof
Hereafter, we fix an arbitrary \(T>0\) and restrict the argument within the finite time horizon [0, T]. Let \({\mathcal {M}}_+([0,1])\) be the space of finite, positive Radon measures on [0, 1], endowed with the weak topology. Define the empirical distribution \(\pi ^N=\pi ^N(t,dx)\) as
where \(\delta _u(dx)\) stands for the Dirac measure at u. Denote by \({\mathcal {D}}={\mathcal {D}}([0,T];{\mathcal {M}}_+([0,1]))\) the space of càdlàg paths on \({\mathcal {M}}_+([0,1])\) endowed with the Skorokhod topology. To prove Theorem 2.8, it suffices to show that the distribution of \(\pi ^N\) on \({\mathcal {D}}\) converges weakly as \(N\rightarrow \infty \) and the limit is concentrated on the single path \(\pi (t,dx)=u(t,x)dx\). However, to formulate the evolution equation (2.7) of u we need a type of convergence that also applies to nonlinear functions. The idea is to introduce the Young measure corresponding to the mesoscopic block average, cf. [12, 13, 22] and [14, Chapter 8].
Let \(\Sigma _T=(0,T)\times (0,1)\). Recall that a Young measure on \(\Sigma _T\) is a measurable map \(\nu :\Sigma _T\rightarrow {\mathcal {P}}({\mathbb {R}})\), where \({\mathcal {P}}({\mathbb {R}})\) is the space of probability measures on \({\mathbb {R}}\) endowed with the topology defined by the weak convergence. Denote by \({\mathcal {Y}}={\mathcal {Y}}(\Sigma _T)\) the set of all Young measures on \(\Sigma _T\), and by \(\nu =\{\nu _{t,x};(t,x)\in \Sigma _T\}\) the element in \({\mathcal {Y}}\). A sequence \(\{\nu ^n;n\ge 1\}\) of Young measures is said to converge to \(\nu \in {\mathcal {Y}}\) if for any bounded and continuous function f on \(\Sigma _T\times {\mathbb {R}}\),
Any measurable function u on \(\Sigma _T\) is naturally viewed as a Young measure:
Denote by \({\mathcal {Y}}_d\) the set of all \(\nu \in {\mathcal {Y}}\) of this kind.
Hereafter, we fix some mesoscopic scale \(K = K (N)\) such that
The existence of such K is guaranteed by (2.17). For \(\eta \in \Omega _N\) and \(i=K\),..., \(N-K\), define the smoothly weighted block average as
Consider the space-time empirical density
where \(\chi _{N,i} (\cdot )\) is the indicator function of the interval \([\tfrac{i}{N} - \tfrac{1}{2N}, \tfrac{i}{N} + \tfrac{1}{2N})\).
Lemma 3.1
(Tightness) Let \({\mathbb {Q}}_N\) be the distribution of \((\pi ^N,\nu ^N)\), where \(\pi ^N\) is defined in (3.1) and \(\nu ^N\) is the Young measure corresponding to \(u^N\) in (3.6) in the sense of (3.3). Then, the sequence of \({\mathbb {Q}}_N\) is tight with respect to the product topology on \(\mathcal D\times {\mathcal {Y}}\).
Let \({\mathbb {Q}}\) be a limit point of \({\mathbb {Q}}_N\). With some abuse of notations, we denote the subsequence converging to \({\mathbb {Q}}\) still by \({\mathbb {Q}}_N\). Below we characterize \({\mathbb {Q}}\) by three propositions.
Proposition 3.2
The following holds for \({\mathbb {Q}}\)-almost every \((\pi ,\nu )\).
- (i):
-
\(\pi (t,dx)=\varpi (t,x)dx\) for every \(t\in [0,T]\) with some \(\varpi (t,\cdot ) \in L^1((0,1))\), and \(t\mapsto \varpi (t,\cdot )\) is a continuous map with respect to the weak topology of \(L^1\).
- (ii):
-
\(\nu _{t,x}([0,1])=1\) for almost all \((t,x)\in \Sigma _T\).
- (iii):
-
\(\varpi (t,x)=\int \lambda \nu _{t,x}(d\lambda )\) for almost all \((t,x)\in \Sigma _T\).
Proposition 3.3
\({\mathbb {Q}}({\mathcal {D}} \times {\mathcal {Y}}_d)=1\), where \({\mathcal {Y}}_d\) is the set of delta-Young measures in (3.3).
To state the last proposition, define the entropy production
for \(\nu \in {\mathcal {Y}}\), \(\varphi \in {\mathcal {C}}^1 ({\mathbb {R}}^2)\) and Lax entropy–flux pair (f, q).
Proposition 3.4
It holds \({\mathbb {Q}}\)-almost surely that
for any Lax entropy–flux pair (f, q) and any \(\varphi \in \mathcal C_c^2([0,T)\times (0,1))\) such that \(\varphi \ge 0\), where \(G(x,\lambda )=V(x)[\lambda -\rho (x)]\).
Remark 3.5
Similarly to Remark 2.2, one can obtain a measure-valued energy bound: it holds \({\mathbb {Q}}\)-almost surely that
This can be derived directly from the microscopic dynamics, see Sect. 6.
Remark 3.6
The arguments we used to prove Lemma 3.1 and Proposition 3.2, 3.3 and 3.4 do apply to all \(V \in {\mathcal {C}}^1((0,1))\), bounded or unbounded. However, only in the non-integrable case are they sufficient to identify the limit equation.
We organize the remaining contents as follows. Lemma 3.1 and Proposition 3.2 are proved in Sect. 4. Proposition 3.3 is proved in Sect. 5 and Proposition 3.4 is proved in Sect. 6.1. With these results, the proofs of Theorem 2.8 and Corollary 2.10 are straightforward and are stated right below. The direct proofs of Proposition 2.11 and (3.9) using the relative entropy method [27] are stated in Sect. 6.2 and 6.3.
Proof of Theorem 2.8 and Corollary 2.10
Recall that \({\mathbb {Q}}\) is a probability measure on \(\mathcal D\times {\mathcal {Y}}\). In view of Proposition 3.2 (i) and 3.3, \(\pi (t)=\varpi (t,x)dx\) and \(\nu _{t,x}=\delta _{u(t,x)}\), \({\mathbb {Q}}\)-almost surely. Proposition 3.2 (iii) then yields that \({\mathbb {Q}}\) is concentrated on the trajectories such that \(\varpi =u\).
To prove Theorem 2.8, we need to show that u, and hence \(\varpi \), is the entropy solution to (2.7) and (2.9). By Proposition 3.2 (ii), \(u(t,x)\in [0,1]\) so that \(u \in L^\infty (\Sigma _T)\). Furthermore, by substituting \(\nu _{t,x}=\delta _{u(t,x)}(d\lambda )\) in Proposition 3.4, we obtain that u satisfies the generalized entropy inequality in Definition 2.1. The proof is then concluded by the uniqueness of the entropy solution, see Proposition 2.4.
Finally, Corollary 2.10 follows directly from the argument above and the sample path regularity of \(\varpi \) obtained in Proposition 3.2 (i). \(\square \)
We close this section with some useful notations. For a function \(\varphi =\varphi (t,x)\), let
Recall the mesoscopic scale \(K=K(N)\) in (3.4). For a sequence \(\{a_i;i=0,\ldots ,N\}\), \({\hat{a}}_{i,K}\) stands for the smoothly weighted average in (3.5). Since K is fixed through the paper, we write \({\hat{a}}_i\) when there is no confusion. We shall frequently use the notions of discrete gradient and Laplacian operators, which are defined as usual:
Notice that \(\Delta =-\nabla \nabla ^*=-\nabla ^*\nabla \).
4 Tightness
Recall that \({\mathbb {Q}}_N\) is the distribution of \((\pi ^N,\nu ^N)\) on \({\mathcal {D}}\times {\mathcal {Y}}\). Here \({\mathcal {D}}=\mathcal D([0,T],{\mathcal {M}}_+([0,1]))\) is the space of càdlàg paths endowed with the Skorokhod topology and \({\mathcal {Y}}=\mathcal Y(\Sigma _T)\) is the space of Young measures endowed with the topology defined by (3.2).
Proof of Lemma 3.1
It suffices to show that both \(\{\pi ^N\}\) and \(\{\nu ^N\}\) are tight. The coordinate \(\nu ^N\) is easy. Since \(u^N\in [0,1]\), for all N we have \({\mathbb {Q}}_N\{\nu ^N\in {\mathcal {Y}}_*\}=1\), where
Because \({\mathcal {Y}}_*\) is compact in \({\mathcal {Y}}\), \(\{\nu ^N\}\) is tight.
For the coordinate \(\pi ^N\), by [14, Chapter 4, Theorem 1.3 & Proposition 1.7], we only need to show that for any \(\psi \in {\mathcal {C}}([0,1])\), some constant \(C_\psi \) and any \(\delta >0\),
where \(\langle \,\cdot ,\cdot \,\rangle \) is the scalar product between \({\mathcal {M}}_+\) and \({\mathcal {C}}([0,1])\). The first one is obvious:
For the second one, choose \(\psi _*\in {\mathcal {C}}_c^2((0,1))\) such that \(\Vert \psi -\psi _*\Vert _{L^1}<4^{-1}\delta \). Note that
uniformly in s, t and all sample paths. Hence, it suffices to show (4.2) with \(\psi \) replaced by \(\psi _*\). Without loss of generality, let \(s<t\). Then,
where \(M_{N,\psi _*}\) is the Dynkin’s martingale. As \(\psi _*\) is compactly supported, \(\eta _0\) or \(\eta _N\) does not appear in \(\langle \pi ^N(r),\psi _* \rangle \), so that
Using the fact that \(\sigma _N \ll N\) and \(\psi _*\in \mathcal C_c^2((0,1))\), \(L_N[\langle \pi ^N,\psi _* \rangle ]\) is uniformly bounded. Therefore, the first term in (4.4) vanishes uniformly when \(|t-s|\rightarrow 0\). We are left with the martingale in (4.4). Dy Dynkin’s formula, the quadratic variation reads
Recall that \(\psi _*\in {\mathcal {C}}_c^2((0,1))\), direct calculation shows that
By (2.17), it is bounded from above by \(CN^{-1}\). Therefore,
We only need to apply Doob’s inequality.
Proof of Proposition 3.2
For (i), notice that (4.3) and the weak convergence yield that for any fixed \(\psi \in {\mathcal {C}}([0,1])\), \({\mathbb {Q}}\{\sup _{t\in [0,T]} |\langle \pi (t),\psi \rangle | \le C\Vert \psi \Vert _{L^1}\}=1\). Since \({\mathcal {C}}([0,1])\) is separable, by standard density argument it holds \({\mathbb {Q}}\)-almost surely that
So \(\pi (t)\) is absolutely continuous with respect to the Lebesgue measure on [0, 1] and can then be written as \(\varpi (t,x)dx\). Moreover, (4.2) assures that \({\mathbb {Q}}\) is concentrated on continuous paths, see [14, Chapter 4, Remark 1.5]. The continuity of \(t \mapsto \varpi (t,\cdot )\) is then proved.
Since (ii) is a direct result from the proof of the tightness of \(\{\nu ^N\}\), we are left with (iii). Pick \(\varphi \in {\mathcal {C}}^1(\Sigma _T)\). From the definition of \(\nu _{t,x}^N\),
where \({{\bar{\varphi }}}_i(t)\) is given by (3.10). By the regularity of \(\varphi \),
Since \(K = K(N) \ll N\), the right-hand side above vanishes as \(N\rightarrow \infty \). From the weak convergence, this yields that for any \(\delta >0\) and fixed \(\varphi \in {\mathcal {C}}^1(\Sigma _T)\),
By choosing a countable and dense subset of \({\mathcal {C}}^1(\Sigma _T)\) and applying (i),
The conclusion in (iii) then follows. \(\square \)
5 Compensated Compactness
Given a Lax entropy–flux pair (f, q), recall the entropy production defined in (3.7). To simplify the notations, we will write \(X(\nu ,\varphi )\) when the choice of (f, q) is clear. Without loss of generality, we also fix \(p=1\) to shorten the formulas.
This section is devoted to the proof of Proposition 3.3. Note that \({\mathcal {Y}}_d\), the subset of delta-type Young measures, is not closed in \({\mathcal {Y}}\), so \({\mathbb {Q}}_N({\mathcal {Y}}_d)=1\) for all N does not guarantee that \({\mathbb {Q}}({\mathcal {Y}}_d)=1\). From [12, Proposition 2.1 & Lemma 5.1], Proposition 3.3 follows from the next result, see also [13, Section 5.6].
Proposition 5.1
Fix an arbitrary Lax entropy–flux pair (f, q). Let \(\varphi =\phi \psi \) with \(\phi \in {\mathcal {C}}_c^\infty (\Sigma _T)\) and \(\psi \in {\mathcal {C}}^\infty ({\mathbb {R}}^2)\). Then, we have the following decomposition:
and there exist random variables \(A_{N,\phi }\), \(B_{N,\phi }\) independent of \(\psi \), such that
Here, \(\Vert \cdot \Vert _{H^1}\) and \(\Vert \cdot \Vert _{L^\infty }\) are the \(H^1\)- and \(L^\infty \)-norm computed on \(\Sigma _T=(0,T)\times (0,1)\).
Remark 5.2
Proposition 5.1 is a microscopic (stochastic) synthesis of the Murat–Tartar theory established in [19, 23]. The role of the function \(\phi \in \mathcal C_c^\infty (\Sigma _T)\) is to localize the estimates in (5.1) and (5.2) away from the boundaries where the potential V is unbounded.
5.1 Basic Decomposition
We first prove a basic decomposition for the microscopic entropy production. Given a bounded function \(\varphi =\varphi (t,x)\), recall the notations \(\varphi _i=\varphi _i(t)\) and \({{\bar{\varphi }}}_i={{\bar{\varphi }}}_i(t)\) defined in (3.10). Denote by \({\hat{\eta _i}}(t)={\hat{\eta }}_{i,K}(t)\), \({\hat{J}}_i(t)={\hat{J}}_{i,K}(t)\) and \({\hat{G}}_i(t)={\hat{G}}_{i,K}(t)\) the smoothly weighted averaged averages introduced in (3.5) of \(\eta _i(t)\), \(J_i(t)=\eta _i(t)(1-\eta _{i+1}(t))\) and \(G_i(t)=V_i(\eta _i(t)-\rho _i)\). We shall abbreviate them to \({\hat{\eta _i}}\), \({\hat{J}}_i\) and \({\hat{G}}_i\) when there is no confusion.
Lemma 5.3
Fix a Lax entropy–flux pair (f, q). For \(\varphi \in \mathcal C^1({\mathbb {R}}^2)\) such that \(\varphi (T,\cdot )=0\),
The terms in (5.3) are defined below. \({\mathcal {A}}_N\), \({\mathcal {S}}_N\) and \({\mathcal {G}}_N\) are given by
\({\mathcal {M}}_N\) is a martingale given by
where \(M_i=M_i(t)\) is the Dynkin’s martingale associated with \(f({\hat{\eta _i}}(t))\), see (5.5) below. Finally, \({\mathcal {E}}_N={\mathcal {E}}_{N,1}+{\mathcal {E}}_{N,2}+{\mathcal {E}}_{N,3}\) is defined through
where \(B_{T,N}\) is the region \(t\in [0,T]\) and \(x\in [0,\tfrac{2K-1}{2N})\cup [1-\tfrac{2K-1}{2N},1]\),
Proof
By the definition of \(\nu ^N\),
By Dynkin’s formula, for \(K \le i \le N-K\),
defines a martingale. Since \(\varphi \) vanishes at \(t=T\), \({\mathcal {M}}_N(\varphi )\) satisfies that
Recall the definition of \(L_N\) in (2.5). Note that, for \(K \le i \le N-K\), \({\hat{\eta }}_i\) does not depend on \(\eta _{0}\) or \(\eta _{N}\), and thus \(L_{\textrm{bd}} {\hat{\eta }}_i = L_\textrm{bd} {\hat{\eta }}_i = 0\). Notice that
Therefore, for \(i=K\),..., \(N-K\), \(L_N[f({\hat{\eta _i}})]\) is equal to
From the above formula of \(L_N[f({\hat{\eta _i}})]\),
We then conclude the proof by inserting (5.6) and (5.7) into the first term on the right-hand side of (5.4). \(\square \)
5.2 Dirichlet Forms
Given a function \(\alpha :[0,1]\rightarrow (0,1)\), denote by \(\nu ^N_{\alpha (\cdot )}\) the product measure on \(\Omega _N\) with marginals
When \(\alpha (\cdot ) \equiv \alpha \) is a constant, we shorten the notation as \(\nu ^N_{\alpha (\cdot )} = \nu ^N_{\alpha }\). Given two probability measures \(\nu \) and \(\mu \) on \(\Omega _N\), let \(f:=\mu /\nu \) be the density function. Define
with \(c_{i,\textrm{G}}\) in (2.2) and \(c_0\), \(c_N\) in (2.3). Note that \(D_-^N\equiv 0\) if \(c_{\textrm{in}}^-=c_\textrm{out}^-=0\), and similarly for \(D_+^N\). Let \(\mu _t^N\) be the distribution of the process at time t. Define
Let \(c_-=c_{\textrm{in}}^-(c_{\textrm{in}}^-+c_\textrm{out}^-)^{-1}\) when at least one of \(c_{\textrm{in}}^-\) and \(c_\textrm{out}^-\) is positive. Define
If \(c_{\textrm{in}}^-=c_{\textrm{out}}^-=0\) we fix \(D_-^N(t)\equiv 0\). Let \(c_+\) and \(D_+^N(t)\) be defined similarly.
Lemma 5.4
For any \(t>0\), there is a constant C independent of N, such that
Proof
Let \(\nu =\nu _\alpha ^N\) with \(\alpha \equiv \tfrac{1}{2}\). For a probability measure \(\mu =f\nu \) on \(\Omega _N\), from the calculation in Appendix A,
For \(s\in [0,t]\), let \(f_s^N:=\mu _s^N/\nu \), then
Applying Lemma 5.5 below, we obtain the estimate
Standard manipulation gives that
is bounded by \(C'N\), so we conclude the proof. \(\square \)
The following a priori bound is used in the previous proof.
Lemma 5.5
Suppose that \(a\in {\mathcal {C}}^1([0,1])\) has Lipschitz continuous derivative. Let \(a_i=a_i^N=a(\tfrac{i}{N})\). Then, there is a constant C independent of N, such that
In particular, the profile \(a(x)=\log [\rho (x)]-\log [1-\rho (x)]\) satisfies the condition in the lemma.
Proof
By Dynkin’s formula,
From the definition of \(L_N\),
where \(j_{i,i+1}=\eta _i(1-\eta _{i+1})+\sigma _N(\eta _i-\eta _{i+1})\). As \(|a_i| \le |a|_\infty \) and \(|\eta _i| \le 1\),
Furthermore, using sum-by-parts formula,
Since \(\sigma _N = o(N)\), one can conclude from the regularity of \(a(\cdot )\). \(\square \)
5.3 Proof of Proposition 5.1
We fix \(\phi \in {\mathcal {C}}_c^\infty (\Sigma _T)\) and prove Proposition 5.1 by estimating each term in the decomposition (5.3) uniformly in \(\psi \). Compared to the proofs of [25, Proposition 6.1] and [24, Lemma 4.3], extra effort is needed to take care of the term with respect to V. In the following contents, C is constant that may depend on (f, q) and T but is independent of \((\phi ,\psi )\), while \(C_\phi \) is constant that also depends on \(\phi \) but is independent of \(\psi \).
We first show that the error term \({\mathcal {E}}_N\) vanishes in the limit.
Lemma 5.6
\({\mathcal {E}}_{N,1}\equiv 0\) for sufficiently large N. Moreover, as \(N \rightarrow \infty \),
- (i):
-
\({\mathcal {E}}_{N,2}\) vanishes uniformly in \(H^{-1} (\Sigma _{T})\), thus satisfies (5.1);
- (ii):
-
\({\mathcal {E}}_{N,3}\) vanishes uniformly in \({\mathcal {M}}(\Sigma _T)\), thus satisfies (5.2).
Proof
For \(\phi \in {\mathcal {C}}_c^1(\Sigma _T)\), there is \(\delta _\phi >0\) such that \(\phi (t,x)=0\) for \(x\notin (\delta _\phi ,1-\delta _\phi )\). Note that \(K \ll N\), so we can find \(N_\phi \) depending only on \(\phi \) such that \(\varphi =\phi \psi =0\) on \(B_{T,N}\) for all \(\psi \) and \(N>N_\phi \). The vanishment of \({\mathcal {E}}_{N,1}\) then follows.
We first prove (i). For \(N>N_\phi \), we can perform summation by parts without generating boundary term:
Note that for some \(\theta _x \in ((2i-1)/2N,x)\),
Thus, by Cauchy-Schwarz inequality and that \(\varphi =\psi \phi \),
To prove (ii), Taylor’s expansion gives that
By the definition of \({\hat{\eta _i}}\) in (3.5), \(|{\hat{\eta }}_i^{j,j+1}-{\hat{\eta _i}}| = K^{-2}\) and \(|{\hat{\eta }}_i^j-{\hat{\eta _i}}| = w_{j-i}\), so
As \(\varphi =\phi \psi \), the definition (3.10) of \({{\bar{\varphi }}}_i\) yields that
Recall that \(\phi =0\) for \(x\notin (\delta _\phi ,1-\delta _\phi )\). Let \(N'_\phi \) be such that \(KN^{-1}<2^{-1}\delta _\phi \) for \(N>N'_\phi \). Then, for every \(N>N'_\phi \), the expression above is bounded by
From the choice of \(w_j\) in (3.5), it is bounded by \(C_\phi K^{-1}\Vert \psi \Vert _{L^\infty }\). Similarly, using Taylor’s expansion and the relation \(f'J'=q'\),
Putting the estimates above together,
The proof is then concluded by the choice of K in (3.4). \(\square \)
Similarly, with the compactness of \(\phi \) we can carry out the estimate for \({\mathcal {G}}_N\).
Lemma 5.7
The functional \({\mathcal {G}}_N\) satisfies (5.2).
Proof
Recall that \(G_i=V_i(\eta _i-\rho _i)\), so \(|G_i| \le C|V(\tfrac{i}{N})|\). Then,
By the argument in Lemma 5.6 (ii), it is bounded by \(C_\phi \Vert \psi \Vert _{L^\infty }\) uniformly in N. \(\square \)
Now, we deal with the martingale term.
Lemma 5.8
The martingale \({\mathcal {M}}_N (\psi )\) satisfies (5.1).
Proof
Since \(\partial _t\varphi =\phi \partial _t\psi +\psi \partial _t\phi \), using Cauchy–Schwarz inequality,
Through Dynkin’s formula, \(M_i\) satisfies that
Similarly to the estimate in Lemma 5.6(ii), \(|{\mathcal {M}}_N(\varphi )| \le a_{N,\phi }\Vert \psi \Vert _{H^1}\) with
The proof is then concluded by the choice of K in (3.4). \(\square \)
To deal with \({\mathcal {A}}_N\) and \({\mathcal {S}}_N\), we need the block estimates stated below. They follow from the upper bound of \(D_\textrm{exc}^N\) in Lemma 5.4 and the logarithmic Sobolev inequality for exclusion process [28]. The proofs are the same as [25, Proposition 6.4 & 6.5] and [24, Proposition 4.6 & 4.7]. For this reason, we omit the details here.
Lemma 5.9
There exists some finite constant C independent of N, such that
Using Lemma 5.9, we can conclude the decompositions of \({\mathcal {A}}_N\) and \({\mathcal {S}}_N\).
Lemma 5.10
Define functionals \({\mathcal {A}}_{N,1}\) and \({\mathcal {S}}_{N,1}\) respectively by
Then, \({\mathcal {A}}_N-{\mathcal {A}}_{N,1}\) and \({\mathcal {S}}_N-{\mathcal {S}}_{N,1}\) satisfy (5.1), while \({\mathcal {A}}_{N,1}\) and \(\mathcal S_{N,1}\) satisfy (5.2).
The proof of Lemma 5.10 follows [25, Lemma 6.6 & 6.7] almost line by line, so we only sketch the difference. It is worth noting that, \({\mathcal {S}}_{N,1}\) turns out to be the only term that survives in the limit, eventually generates the non-zero macroscopic entropy in (2.10).
Proof
We first treat \({\mathcal {A}}_N\). Since \(\varphi =\phi \psi \) with \(\phi \in {\mathcal {C}}_c^\infty (\Sigma _T)\) being fixed, similarly to the proof of Lemma 5.6 (i), we have for \(N>N_\phi \) that
Applying Cauchy–Schwarz inequality and Lemma 5.9, we obtain that \(|({\mathcal {A}}_N-{\mathcal {A}}_{N,1})(\varphi )| \le a_N\Vert \varphi \Vert _{H^1}\) and \(|{\mathcal {A}}_{N,1}(\varphi )| \le b_N\Vert \varphi \Vert _{L^\infty }\), where \((a_N,b_N)\) are random variables such that
Noting that \(\Vert \varphi \Vert _{H^1} \le C_\phi \Vert \psi \Vert _{H^1}\), \(\Vert \varphi \Vert _{L^\infty } \le C_\phi \Vert \psi \Vert _{L^\infty }\), the conclusion follows from (3.4).
The proof for \({\mathcal {S}}_N\) is similar. For \(N>N_\phi \),
By Cauchy–Schwarz inequality and Lemma 5.9, \(|(\mathcal S_N-{\mathcal {S}}_{N,1})(\varphi )| \le a'_N\Vert \varphi \Vert _{H^1}\) and \(|{\mathcal {S}}_{N,1}(\varphi )| \le b'_N\Vert \varphi \Vert _{L^\infty }\) with random variables \((a'_N,b'_N)\) satisfying
The conclusion follows similarly. \(\square \)
6 Measure-Valued Entropy Solution
We prove that under \({\mathbb {Q}}\), \(\nu \) satisfies (3.8) with probability 1. We call such a Young measure a measure-valued entropy solution to the initial-boundary value problem (2.7) and (2.9). Thanks to Proposition 3.3, \(\nu \) is essentially the entropy solution.
6.1 Proof of (3.8)
Fix an arbitrary Lax entropy–flux pair (f, q) and \(\varphi \in {\mathcal {C}}_c^2([0,T)\times (0,1))\) such that \(\varphi \ge 0\). As in Sect. 5, we shall let \(N\rightarrow \infty \) and examine the limit of each term in the decomposition (5.3). The main difference is that the proof of Proposition 5.1 requires uniform estimate in the test function, which is no more necessary here.
First, by Lemma 5.6, 5.8 and 5.10, for any \(\delta >0\),
Meanwhile, the convexity of f ensures that \(\nabla {\hat{\eta _i\nabla }} f'({\hat{\eta _i}}) \ge 0\), so \({\mathcal {S}}_{N,1}(\varphi )\) given by (5.18) is non-positive everywhere. Therefore, (3.8) holds \({\mathbb {Q}}\)-almost surely if we can show that
for any \(\delta >0\), where \(G(x,\lambda )=V(x)(\lambda -\rho (x))\). Recall the definition of \({\mathcal {G}}_N(\varphi )\) in Lemma 5.3. As \(\varphi \) is compactly supported, for sufficiently large N we have
where \({\hat{G}}_i\) is the smoothly weighted average of \(G_i=V(\tfrac{i}{N})(\eta _i-\rho (\tfrac{i}{N}))\). Straightforward computation shows that
which vanishes uniformly as \(N\rightarrow \infty \). We can then conclude (6.1).
6.2 Direct Proof of (3.9)
Notice that (3.9) holds if for some constant \(C_0\),
where \({\bar{u}}={\bar{u}}(t,x):=\int _{\mathbb {R}}\lambda \nu _{t,x}(d\lambda )\) is a measurable function on \(\Sigma _T\).
Let \(\{g^j;j \ge 1\}\) be a countable subset of \({\mathcal {C}}^\infty _c ({\mathbb {R}}\times (0,1))\) which is dense in both the \(L^1\) and the \(L^2\) norm induced by V, i.e., for any \(g\in {\mathcal {C}}^\infty _c ({\mathbb {R}}\times (0,1))\), there are \(g^{j_n}\), \(n\ge 1\), such that
For \(\ell \ge 1\), consider the functional \(\Phi _\ell :{\mathcal {Y}}\rightarrow {\mathbb {R}}\) defined as
Note that \(\Phi _\ell \) is continuous for each fixed \(\ell \). Lemma 6.1 below together with the weak convergence of \({\mathbb {Q}}_N\) then shows that there is a constant \(C_1\) independent of \(\ell \), such that
Taking \(\ell \rightarrow \infty \) and applying the monotone convergence theorem,
The condition (6.2) then follows from the dense property of \(\{g^j;j\ge 1\}\).
Lemma 6.1
There exist constants \(C_0\) and \(C_1\) such that, for each \(\ell \ge 1\),
Proof
Only in this proof, to shorten the formulas we denote
Since \(g^j\) is compactly supported, for N sufficiently large,
To conclude the proof, it suffices to prove that
We begin with (6.4), which is completely deterministic. For some fixed \(j\le \ell \), using the fact that \(g^j\) is compactly supported, we only need to show that
This follows as \(\rho \) is continuous and \({\hat{\rho _i}}\) is the smoothly weighted average of \(\rho (\tfrac{i}{N})\).
Now we prove (6.5). First note that it suffices to prove it with K replaced by any other mesoscopic scale \(n=n(N)\) such that \(K \le n = o(N)\), since
The choice of n is specified below in Lemma 6.2. Recall that \(\nu _{\rho (\cdot )}^N\) is the product measure on \(\Omega _N\) associated with the profile \(\rho \). From the entropy inequality, the expectation in (6.5) is bounded by
Due to (2.4), the relative entropy is bounded by CN, so the first term is uniformly bounded. Also notice that for any random variables \(X_1\),... \(X_\ell \),
so we only need to find universal constants \(C_0\) and \(C_1\), such that
for all \(g\in {\mathcal {C}}_c^\infty ({\mathbb {R}}\times (0,1))\). By the Feynman–Kac formula (see, e.g., [2, Lemma 7.3]), the left-hand side of (6.6) is bounded from above by
where the supremum is taken over all \(\nu _{\rho (\cdot )}^N\)-densities. For each j, performing the change of variables \(\eta \mapsto \eta ^j\),
where the last line follows from the equality \((1- \eta _j - \rho _j) \nu ^N_{\rho (\cdot )} (\eta ^j) = - (\eta _j - \rho _j) \nu ^N_{\rho (\cdot )} (\eta )\). Thus, the first term inside the supremum in (6.7) reads
Using Cauchy–Schwarz inequality, we can bound it from above by \(\mathcal I_1+{\mathcal {I}}_2\), where
where \(c_{i,\textrm{G}}(\eta )=V_i[\rho _i(1-\eta _i)+\eta _i(1-\rho _i)]>0\) due to (2.4). Recall the Dirichlet form \(D^N_\textrm{G}\) defined in (5.9). Since \(\sum _{|j|<K} w_j=1\),
where \(\mu :=f\nu _{\rho (\cdot )}^N\). To estimate \({\mathcal {I}}_2\), notice that
From (2.4), Lemma 6.2 below and Cauchy–Schwarz inequality,
with some constant C independent of (i, j, g). Therefore, \({\mathcal {I}}_2 \le C_0N\Vert g\Vert _V^2\). Putting the estimates for \({\mathcal {I}}_1\), \({\mathcal {I}}_2\) into (6.7), we see that (6.6) holds if we can show
for any N, any \(\nu _{\rho (\cdot )}^N\)-density f and \(\mu =f\nu _{\rho (\cdot )}^N\)
The proof of (6.8) is standard. Since \(\nu _{\rho (\cdot )}^N\) is reversible for \(L_{\textrm{G}}\),
In Appendix A, we prove that (cf. (5.12) and (5.14))
Since the Dirichlet forms are non-negative, (6.8) follows. \(\square \)
Lemma 6.2
Suppose that \(V\in {\mathcal {C}}^1((0,1))\) and \(\inf _{(0,1)} V>0\). Define \(n=n(N)\) as
Then, \(n=o(N)\) as \(N\rightarrow \infty \), and there is a constant \(C=C(V)\) such that
Proof
We first prove that for any \(\varepsilon >0\), \(n\le \varepsilon N\) for sufficiently large N. Indeed, let \(N_\varepsilon \) be such that \(KN^{-1}<2^{-1}\varepsilon \) for all \(N>N_\varepsilon \). Then, if \(N>N_\varepsilon \),
We can further choose \(N_\varepsilon \) such that \(KN^{-1}<C_\varepsilon ^{-1}\) for all \(N>N_\varepsilon \), then \(n\le \varepsilon N\).
For the second criterion, suppose that \(|V(\tfrac{i+j}{N})|\) takes the minimum value for \(|j|<K\) at some \(j_{N,i}\). Then, for each \(i=n\),..., \(N-n\),
Therefore, for each i and j,
The second criterion then follows from (2.4). \(\square \)
6.3 Direct Proof of Proposition 2.11
By the continuity of \(\rho (\cdot )\),
Thus, we only need to show for fixed \(\varepsilon >0\) that
Take
Let \(\mu ^{(\varepsilon N)} (s)\) denote the distribution of \(\{\eta _1 (s), \eta _2 (s), \ldots , \eta _{\varepsilon N} (s)\}\). By entropy inequality,
We first bound \(\textrm{II}\), which is simpler. Note that we could first remove the absolute value inside the exponential. Since \(\nu ^N_{\rho (\cdot )}\) is a product measure, and by Taylor’s expansion, there exists some constant C such that
which converges to zero as \(\varepsilon \rightarrow 0\) by (2.20).
For \(\textrm{I}\), we consider the following Markov chain \(X(t):= \{X_i(t)\}_{1\le i \le \varepsilon N}\), where \(\{X_i (t)\}\), \(1 \le i \le \varepsilon N\), are independent \(\{0,1\}\)-valued Markov chains, and the transition rates for \(X_i (t)\) are given by
Since \(\rho (\cdot )\) is bounded away from zero and one, the logarithmic Sobolev constant for the chain \(X_i (t)\) is of order \(V_i\). By [10, Lemma 3.2], the logarithmic Sobolev constant for the chain X(t) has order
Therefore,
By Lemma 5.4, \(\int _0^t D^N_{\textrm{G}} (s) ds \le C\). Thus,
which also converges to zero as \(\varepsilon \rightarrow 0\).
Data Availability
The authors declare that all data supporting this article are available within the article.
References
Bahadoran, C.: Hydrodynamics and hydrostatics for a class of asymmetric particle systems with open boundaries. Commun. Math. Phys. 310, 1–24 (2012). https://doi.org/10.1007/s00220-011-1395-6
Baldasso, R., Menezes, O., Neumann, A., Souza, R.: Exclusion process with slow boundary. J. Stat. Phys. 167, 1112–1142 (2017). https://doi.org/10.1007/s10955-017-1763-5
Bardos, C., LeRoux, A.Y., Nédélec, J.C.: First order quasilinear equations with boundary conditions. Commun. Partial Differ. Equ. 4(9), 1017–1034 (1979). https://doi.org/10.1080/03605307908820117
Bernardin, C., Cardoso, P., Gonçalves, P., Scotta, S.: Hydrodynamic limit for a boundary driven super-diffusive symmetric exclusion. Stoch. Process. Appl. 165, 43–95 (2023). https://doi.org/10.1016/j.spa.2023.08.002
Bernardin, C., Gonçalves, P., Jiménez-Oviedo, B.: Slow to fast infinitely extended reservoirs for the symmetric exclusion process with long jumps. Markov Process. Relat. Fields 25, 217–274 (2019)
Bernardin, C., Gonçalves, P., Jiménez-Oviedo, B.: A microscopic model for a one parameter class of fractional Laplacians with Dirichlet boundary conditions. Arch. Ration. Mech. Anal. 239, 1–48 (2021). https://doi.org/10.1007/s00205-020-01549-9
De Masi, A., Marchesani, S., Olla, S., Xu, L.: Quasi-static limit for the asymmetric simple exclusion. Probab. Theory Relat. Fields 183, 1075–1117 (2022). https://doi.org/10.1007/s00440-022-01140-1
De Masi, A., Presutti, E., Tsagkarogiannis, D., Vares, M.E.: Current reservoirs in the simple exclusion process. J. Stat. Phys. 144, 1151–1170 (2011). https://doi.org/10.1007/s10955-011-0326-4
Derrida, B., Evans, M.R., Hakim, V., Pasquier, V.: Exact solution of a 1D asymmetric exclusion model using a matrix formulation. J. Phys. A 26(7), 1493–1517 (1993). https://doi.org/10.1088/0305-4470/26/7/011
Diaconis, P., Saloff-Coste, L.: Logarithmic Sobolev inequalities for finite Markov chains. Ann. Appl. Probab. 6(3), 695–750 (1996). https://doi.org/10.1214/aoap/1034968224
Erignoux, C., Simon, M., Zhao, L.: Asymmetric attractive zero-range processes with particle destruction at the origin. Stoch. Process. Appl. 159, 1–33 (2023). https://doi.org/10.1016/j.spa.2023.01.015
Fritz, J.: Entropy pairs and compensated compactness for weakly asymmetric systems. In: Funaki, T., Osada, H. (eds.) Stochastic Analysis on Large Scale Interacting Systems. Advanced Studies in Pure Mathematics, vol. 39, pp. 143–171. Mathematical Society of Japan, Kyoto (2004)
Fritz, J., Tóth, B.: Derivation of the Leroux system as the hydrodynamic limit of a two-component lattice gas. Commun. Math. Phys. 249, 1–27 (2004). https://doi.org/10.1007/s00220-004-1103-x
Kipnis, C., Landim, C.: Scaling Limits of Interacting Particle Systems, vol. 320. Springer Science & Business Media, Berlin (2013)
Landim, C., Milanés, A., Olla, S.: Stationary and nonequilibrium fluctuations in boundary driven exclusion processes. Markov Process. Relat. Fields 14(2), 165–184 (2008)
Liu, T.-P.: Hyperbolic conservation laws with relaxation. Commun. Math. Phys. 108, 153–175 (1987). https://doi.org/10.1007/BF01210707
Marchesani, S., Olla, S., Xu, L.: Quasi-static limit for a hyperbolic conservation law. Nonlinear Differ. Equ. Appl. NoDEA 28, 53 (2021). https://doi.org/10.1007/s00030-021-00716-5
Martin, S.: First order quasilinear equations with boundary conditions in the \(L^\infty \) framework. J. Differ. Equ. 236(2), 375–406 (2007). https://doi.org/10.1016/j.jde.2007.02.007
Murat, F.: Compacitè par compensation. Ann. Sci. Norm. Sup. Pisa 5(3), 489–507 (1978)
Otto, F.: Initial-boundary value problem for a scalar conservation law. C. R. Acad. Sci. Paris 322(1), 729–734 (1996)
Popkov, V., Schütz, G.M.: Steady-state selection in driven diffusive systems with open boundaries. Europhys. Lett. 48(3), 257–263 (1999). https://doi.org/10.1209/epl/i1999-00474-0
Rezakhanlou, F.: Hydrodynamic limit for attractive particle systems on \({\mathbb{Z} }^{d}\). Commun. Math. Phys. 140(3), 417–448 (1991). https://doi.org/10.1007/BF02099130
Tartar, L.: Compensated compactness and applications to partial differential equations. In: Knops, R. (ed.) Nonlinear Analysis and Mechanics: Heriot-Watt Symposium, Pitman Research Notes in Mathematics, vol. 39, pp. 136–212. Pitman, London (1979)
Xu, L.: Hydrodynamics for one-dimensional ASEP in contact with a class of reservoirs. J. Stat. Phys. 189, 1 (2022). https://doi.org/10.1007/s10955-022-02963-x
Xu, L.: Hydrodynamic limit for asymmetric simple exclusion with accelerated boundaries. (2021). To appear in Ann. Inst. Henri Poincaré Probab. Statist.https://arxiv.org/abs/2108.09345
Xu, L.: Scalar conservation law in a bounded domain with strong source at boundary. Nonlinear Differ. Equ. Appl. NoDEA 31, 66 (2024). https://doi.org/10.1007/s00030-024-00959-y
Yau, H.-T.: Relative entropy and hydrodynamics of Ginzburg–Landau models. Lett. Math. Phys. 22, 63–80 (1991). https://doi.org/10.1007/BF00400379
Yau, H.-T.: Logarithmic Sobolev inequality for generalized simple exclusion processes. Probab. Theory Relat. Fields 109, 507–538 (1997). https://doi.org/10.1007/s004400050140
Acknowledgements
The authors specially thank an anonymous referee whose earnest work helped to improve the paper. Linjie Zhao is supported by National Natural Science Foundation of China (Grant Number 12371142) and the financial support from the Fundamental Research Funds for the Central Universities in China.
Funding
Open access funding provided by Gran Sasso Science Institute - GSSI within the CRUI-CARE Agreement.
Author information
Authors and Affiliations
Corresponding author
Ethics declarations
Conflict of interest
All authors have no Conflict of interest.
Additional information
Communicated by Stefano Olla.
Publisher's Note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
Appendix A: Computations Concerning the Dirichlet Forms
Appendix A: Computations Concerning the Dirichlet Forms
Here we collect some fundamental estimates of the Dirichlet forms in (5.8)–(5.11).
1.1 A.1: Proof of (5.12)
From the definition of \(L_{\textrm{exc}}\),
where the operators \(S_{\textrm{exc}}\) and \(A_{\textrm{exc}}\) are respectively given by
Recall that \(\nu =\nu _{\frac{1}{2}}^N\) and f is a \(\nu \)-density. By the basic inequality \(x\log (y/x) \le 2\sqrt{x}(\sqrt{y}-\sqrt{x})\) for any x, \(y\ge 0\),
We compute \(S_{\textrm{exc}}\) and \(A_{\textrm{exc}}\) respectively. Since \(\nu \) is reversible for \(S_{\textrm{exc}}\),
where \(\mu =f\nu \). For \(A_{\textrm{exc}}\), we only need to observe that
The estimate is then concluded.
1.2 A.2: Proof of (5.13)
Let \(f_*\) be the density of \(\mu \) with respect to \(\nu _{\rho (\cdot )}^N\), then
Therefore, \(\langle f,L_{\textrm{G}}[\log f] \rangle _\nu \) equals to
Since \(\nu _{\rho (\cdot )}^N\) is reversible with respect to \(L_\textrm{G}\), similarly as in Appendix A.1,
The estimate (5.13) then holds since \(c_{i,\textrm{G}}(\eta )(1-2\eta _i) = V_i(\rho _i-\eta _i)\).
1.3 A.3: Proof of (5.14)
As in Appendix A.1, we begin with
We compute the terms associated with \(\eta _0\) as an example. Those associated with \(\eta _N\) follow the same argument. Notice that
The conclusion then follows since \(\nu (\eta ^0)=\nu (\eta )\) and \(c_0(\eta ^0)-c_0(\eta ) = (c_{\textrm{out}}^--c_{\textrm{in}}^-)(1-2\eta _0)\).
1.4 A.4: Proof of (6.9)
Let f be a \(\nu _{\rho (\cdot )}^N\)-density function and denote \(\mu =f\nu _{\rho (\cdot )}^N\). Observe that the main difference from (5.12) is that the reference measure \(\nu _{\rho (\cdot )}^N\) is not reversible. Recall the decomposition (A.1) and we estimate \(S_{\textrm{exc}}\) and \(A_{\textrm{exc}}\) respectively.
We begin with the symmetric part \(S_{\textrm{exc}}\). Let \(g:=\sqrt{f}\). By dividing \(g(\eta )\) into \(2^{-1}(g(\eta )-g(\eta ^{i,i+1}))\) and \(2^{-1}(g(\eta )+g(\eta ^{i,i+1}))\), we obtain
Applying the change of variable \(\eta ^{i,i+1}\rightarrow \eta \), the second term becomes
Since \(\rho \in {\mathcal {C}}^1([0,1];(0,1))\), with a constant C independent of (i, N) we have
Since \(g^2\nu _{\rho (\cdot )}^N=\mu \) is a probability measure, we have
Now we treat the antisymmetric part \(A_{\textrm{exc}}\). First observe that
As \(|\eta _{i+1}-\eta _i|\le 1\), the first term is bounded by \(D^N_\textrm{exc}(\mu ;\nu _{\rho (\cdot )}^N)\). The second term reads
Since \(f(\eta )\nu _{\rho (\cdot )}^N(\eta )=\mu (\eta )\), its modulus is bounded by
which is uniformly bounded due to (A.2). Therefore,
Putting the two estimates into (A.1), we can conclude since \(\sigma _N \ll N\).
1.5 A.5: Proof of (6.10)
Repeating the argument in the proof of (6.9) with \(\nu \) replaced by \(\nu _{\rho (\cdot )}^N\), we only need to bound
It is uniformly bounded since
Rights and permissions
Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article's Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article's Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creativecommons.org/licenses/by/4.0/.
About this article
Cite this article
Xu, L., Zhao, L. Hydrodynamics for Asymmetric Simple Exclusion on a Finite Segment with Glauber-Type Source. J Stat Phys 191, 78 (2024). https://doi.org/10.1007/s10955-024-03297-6
Received:
Accepted:
Published:
DOI: https://doi.org/10.1007/s10955-024-03297-6