Derivation of coupled KPZ-Burgers equation from multi-species zero-range processes (Q2240874)

The paper under review studies scaling limits of the empirical mass fluctuation fields of weakly asymmetric multi-species zero-range processes on a sequence of tori \(T_N=\{0, 1, \cdots, N-1\}\) as an approximation of \(\mathbb Z\). It is natural to analyze the connections between the fluctuations when there is more than one conservation law and coupled KPZ-Burgers systems. The main result of the paper, Theorem 3.4, shows that the mass fluctuation limit of the multi-species process satisfies a type of martingale problem for coupled KPZ-Burgers equation provided the characteristic speed is same up to leading order of each component. The frame condition (FC) is on the density of the stationary state, but not every stationary state satisfies the FC. Determining when the FC holds is the first microscopic derivation of a singular coupled SPDE system studied in this paper. Section 2 gives the model with the weakly asymmetric \(n\)-species zero-range process. The weakly asymmetric process \(\{\alpha_t^N = (\alpha_t^{1, N}, \cdots, \alpha_t^{n, N})\}_{t\ge 0}\in \Omega^n\) for \(\Omega = \mathbb Z_+^{T_N}\) the configuration space of particles of a single species and \(\mathbb Z_+=\{0, 1, 2, \cdots\}\), corresponds to the case where the jump probabilities \( (p_i: p_i (\pm 1) = 1/2 \pm c_i^N, p_I(z)=0, z\neq \pm 1)_{i=1}^n\) for \(c_i^N = O(1/N^{\gamma})\) with \(\gamma >0\) as \(N\to \infty\). When \(\gamma = 1\), it gives a hydrodynamic limit and linear fluctuation in the Euler time scale; when \(\gamma = 1/2\) it provides the second order KPZ-Burgers fluctuations in the diffusive time-scale. There are four assumptions on the weakly asymmetric \(n\)-species zero-range process: \begin{itemize} \item[(1)] (ND) Nondegeneracy condition on the jump rate \(g_i^* >0\) for \(i=1, \cdots, n\); \item[(2)] (INV) The jump rate \((g_i)_{i=1}^n\) is compatible and (ORI) \(\phi_* >0\); \item[(3)] (SG) for sufficient mixing for the Boltzmann-Gibbs principle; and \item[(4)] (FC) The frame condition for \[ \lambda = \lambda (a_0) = - \partial_{a^i}(\tilde{g}_i(a))|_{a_0}, \ \ \ \partial_{a^j}(\tilde{g}_i(a))|_{a_0}=0, i\neq j. \] \end{itemize} When the model is on \(\mathbb Z\), an additional (LG) linear growth assumption for an upper bound Lipschitz criterion is needed to construct the process. Section 3 states the main result Theorem 3.1 with limiting process in the uniform topology, which solves uniquely the SPDE \[ \partial_t Y_t = \frac{1}{2}Q(a_0) \Delta Y_t - 2CQ(a_0)\nabla Y_t + q(a_0) \nabla \dot{W}_t. \] The SPDE is a cross-diffusion system, hence its well-posedness is nontrivial. By the limiting procedure, this reduces to show the well-posedness and invariance of the distribution of \(\dot{W}_0\) the Gaussian distribution taking values in \({\mathcal D}^{'} (T^n)\), in Lemma 3.2. With the frame condition, Proposition 3.3 shows that this assumption is equivalent to: the covariance matrix is diagonal. Then the limits \(Y_t\) of \(\{Y_t^N\}_{N\ge 1}\) satisfy a type of ill-posed KPZ-Burgers equation (3.4) (see the paper), and Theorem 3.4 shows that \(Y_t\) is the unique multi-species stationary energy solution if \(Y_t\) is a spatial white noise with \(L^2\) energy, and there is a Brownian motion with respect to the filtration generated by \(Y_t\) and proper quadratic and cross variations provided \(c_i^N= \frac{c}{\sqrt{N}} + \frac{c_i}{N}\). Section 4 devotes to the proofs of Theorem 3.1 and Theorem 3.4 by following the hydrodynamics scheme of \textit{P. Gonçalves} and \textit{M. Jara} [Arch. Ration. Mech. Anal. 212, No. 2, 597--644 (2014; Zbl 1293.35336)] with some new properties to the multi-species context. The general generators are computed in Section 4.1 (By Dynkin's formula, the process \(M_t^N\) is martingale, and decomposed by \(M_t^N = Y_t^N - Y_0^N - I_t^N - B_t^N - K_t^N\)), a general Boltzmann-Gibbs principle is stated in Section 4.2 (for generic density \(a_0\), Theorem 4.1 for \(B_t^{i, N}\) part estimate), tightness of the processes in Theorem 3.4 is proved in Section 4.3 (Proposition 4.2 is proved by Mitoma's criterion and Doob inequality through the Kolmogorov-Centsov criterion), identity some features of the limits in Section 4.4 (use the decomposition of \(M_t^N\) to study the limits of each term in Proposition 4.3) and completes the proof of Theorem 3.4 in Section 4.5 (follows by the decomposition and Proposition 4.3 and tightness in the uniform topology, and the unique characterization of the multi-species energy solution). Section 5 provides a proof of Theorem 4.1 for the Boltzmann-Gibbs principle with most technique lemmas in estimations. Section 6 shows that the coupled KPZ-Burgers equation obtained in the limit satisfies the trilinear condition for the coupling constants \(\Gamma_{jl}^i\). Section 7 focuses on a special case of multi-colored zero-range processes with a series of computations in lemmas, and the last section 8 constructs a model which satisfies all necessary conditions (ND), (INV), (ORI) and (SG) and (FC) to illustrate that the KPZ-Burgers system cannot be fully decoupled.