Uniform stability estimate for the Vlasov-Poisson-Boltzmann system (Q2229364)

In this paper the author studies the uniform stability and convergence to the steady state of the Vlasov-Poisson-Boltzmann system of equations. The Cauchy problem under consideration is \[ \partial_t F + v \cdot \nabla_x F + \nabla_x \Phi \cdot \nabla_v F = Q(F,F),\; F(0,x,v) = F_0(x,v), \] \[ \Delta_x \Phi = \int_{R^3} F\; dv - 1. \] Here \(x=(x_1,x_2,x_3),\; v=(v_1,v_2,v_3)\). The function \(F(t,x,v)\) is a number density for charged particles in the presence of a fixed background of opposite charge. The function \(\Phi(t,x)\) represents the internally consistent electrostatic field. The bilinear collision operator \(Q(F,G)\) is for hard sphere interaction and has the form \[ Q(F,G) = \int_{R^3} \int_{S^2}|(v-v_*) \cdot \omega| (G'_*F'-G_*F)\; dv_* d\omega, \] where \(G_*,F\) and \(G_*',F'\) are function evaluate before and after collisions.The analysis of (1),(2) is restricted to fluctuations around the steady state Maxwellian distribution given by \[ \mu(v) = (2\pi)^{-3/2}e^{-|v|^2/2}. \] Accordingly a perturbation function is defined by \(F(t,x,v) = \mu + \sqrt{\mu} f\). In terms of \(f(t,x,v)\) the system (1),(2) then becomes \[ \partial_t f + v \cdot \nabla_x f + \nabla_x \Phi \cdot \nabla_v f - \frac{1}{2}v \cdot \nabla_x \Phi f - \nabla_x \Phi \cdot v \sqrt{\mu} = Lf + \Gamma(f,f), \] \[ \Delta_x \Phi = \int_{R^3} \sqrt{u} f\; dv - 1, \;\; f(0,x,v) = f_0(x,v). \] Here \(L\) and \(\Gamma\) are respectively the linear and nonlinear collision terms given by \[ Lf = \mu^{-1/2}Q(\mu,\sqrt{\mu}f) + \mu^{-1/2}Q(\sqrt{\mu}f, \mu), \] \[ \Gamma(f,g) = \mu^{-1/2}Q(\sqrt{\mu}f,\sqrt{\mu} g), \] The null space of the operator \(L\) is denoted \(\mathcal N\), and the operator \(\mathbf P\) is the orthogonal projection of \(L_2(R_v^3)\) to \(\mathcal N\). The function \(f(t,x,v)\) is given the decomposition \(f(t,x,v) = {\mathbf P}f(t,x,v) + \{{\mathbf I} - {\mathbf P} \} f(t,x,v)\) where \({\mathbf P} f\) and \(\{{\mathbf I} - {\mathbf P} \} f\) are called the macroscopic and microscopic components of \(f\) respectively. The main results of this paper are stated in two theorems. The first theorem (Theorem 1.1 in the paper) has been proved in some previous papers that are cited but is stated in the present paper. The statement of the theorem is that an energy functional \({\mathcal E}_N(\cdot)\) and a dissipation rate \({\mathcal D}_N(\cdot)\) are defined in terms of \(L_2\) derivatives of the functions \(f,\; \nabla_x \Phi,\;{\mathbf P}f,\; {\{\mathbf I - \mathbf P \}} f\). Here \(N\) designates the degree of the \(L_2\) derivatives. Then for \(N \ge 4\) and \({\mathcal E}_N(f)(0) \le \delta\) sufficiently small the Cauchy problem (3),(4) has a unique solution, and \[\frac{d}{dt} {\mathcal E}_N(f)(t) + \lambda {\mathcal D}_N(f)(t) \le 0,\;\; t \ge 0. \] With some further assumptions on \(f_0\) it is stated that \(f(t)\) converges to zero according to \[ ||f(t)||_{H^N} \le C\delta (1+t)^{-3/4}. \] The main new result of the present paper is given by the second theorem (Theorem 1.2 in the paper) and provides a uniform stability estimate on the solution to (3),(4). In this theorem it is proved that if \(\max \{{\mathcal E}_N(f')(0),{\mathcal E}_N(f'')(0) \} \le \delta\) and with some additional conditions on \(f_0',f_0''\) then for \(N \ge 5\) \[ {\mathcal E}_{N-1}(f'-f'')(t) \le C{\mathcal E}_N(f'-f'')(0), \;\; t \ge 0. \] A significant part of the analysis in the paper is to construct a Kawashima compensating function used to establish estimates on the macroscopic part of the solution, \(f\). Then with some additional energy estimates and the use of Theorem 1.1 the proof of Theorem 1.2 is established.