Homogenization of linear transport equations. A new approach (Q2305470)

The author proves a homogenization result for the linear transport equation \( \frac{\partial u_{\varepsilon }}{\partial t}-b_{\varepsilon }(x)\cdot \nabla _{x}u_{\varepsilon }=0\) posed in \((0,T)\times \mathbb{R}^{N}\), with the initial condition \(u_{\varepsilon }(0,\cdot )=u_{\varepsilon }^{0}\) in \( \mathbb{R}^{N}\), where \(u_{\varepsilon }^{0}\) is a bounded sequence in \( L^{p}(\mathbb{R}^{N})\), \(N\geq 2\), and \(b_{\varepsilon }\) is a vector field in \(C_{b}^{0}(\mathbb{R}^{N})^{N}\cap W_{\mathrm{loc}}^{1,q}(\mathbb{R}^{N})^{N}\). The author assumes the existence of a positive function \(\sigma _{\varepsilon }\) in \(C^{0}(\mathbb{R}^{N})\cap W_{\mathrm{loc}}^{1,q}(\mathbb{R}^{N})\) satisfying \(c^{-1}\leq \sigma _{\varepsilon }\leq c\) for some constant \(c>1\) and \(\sigma _{\varepsilon }b_{\varepsilon }=R_{\perp }\nabla w_{\varepsilon }^{2}\) if \(N=2\) and \(\sigma _{\varepsilon }b_{\varepsilon }=\nabla w_{\varepsilon }^{2}\times \ldots \times \nabla w_{\varepsilon }^{N}\) if \( N\geq 3\), where \(W_{\varepsilon }=(w_{\varepsilon }^{1},\ldots ,w_{\varepsilon }^{N})\) is a sequence of vector fields in \(C^{1}(\mathbb{R} ^{N})^{N}\) which is uniformly proper in the sense that for any compact set \(K \) of \(\mathbb{R}^{N}\) there exists a compact subset \(K^{\prime }\) of \( \mathbb{R}^{N}\) satisfying \(W_{\varepsilon }^{-1}(K)\subset K^{\prime }\) for any small enough \(\varepsilon >0\) and such that \(W_{\varepsilon }\rightarrow W\) in \(C_{\mathrm{loc}}^{0}(\mathbb{R}^{N})^{N}\) and \(W_{\varepsilon }\rightharpoonup W\) in \(W_{\mathrm{loc}}^{1,N}(\mathbb{R}^{N})^{N}\) for some\ \(W\in C^{1}(\mathbb{R} ^{N})^{N}\). The author also assumes that for \(p\in (1,\infty )\) with conjugate exponent \(q\), there exists a positive function \(\theta _{0}\) in \( C^{0}(\mathbb{R}^{N})\) such that \(\theta _{\varepsilon }:=b_{\varepsilon }\cdot \nabla w_{\varepsilon }^{1}>0\) in \(\mathbb{R}^{N}\) and \(\theta _{\varepsilon }\rightarrow \theta _{0}>0\) strongly in \(L^{q}(\mathbb{R}^{N})\) and that either \(\mathrm{div}\, b_{\varepsilon }\) is bounded in \(\mathbb{R}^{N}\) or \(b_{\varepsilon }\in C_{b}^{1}(\mathbb{R}^{N})^{N}\), \(\sigma _{\varepsilon }\in C^{1}(\mathbb{R}^{N})\) and \(u_{\varepsilon }^{0}\in C^{1}( \mathbb{R}^{N})\). The main result of the paper proves that if \( u_{\varepsilon }\) is the solution to the above transport equation the function \(v_{\varepsilon }:=\sigma _{\varepsilon }u_{\varepsilon }\) converges, up to a subsequence, weakly in \(L^{1}(0;T;L^{p}(\mathbb{R}^{N}))\) to a solution \(v\) to the transport equation \(\frac{\partial v}{\partial t} -\xi _{0}\cdot \nabla _{x}(\frac{v}{\sigma _{0}})=0\) in \((0,T)\times \mathbb{ R}^{N}\), with the initial condition \(v(0,\cdot )=v^{0}\) in \(\mathbb{R}^{N}\), where \(\xi _{0}=Cof(DW)e_{1}\in C^{0}(\mathbb{R}^{N})^{N}\) (\(Cof\) denotes the cofactors matrix), \(\sigma _{\varepsilon }b_{\varepsilon }\rightharpoonup \xi _{0}\) in \(L_{\mathrm{loc}}^{N/(N-1)}(\mathbb{R}^{N})^{N}\), \( \sigma _{\varepsilon }\rightharpoonup \sigma _{0}\) in \(L^{\infty }(\mathbb{R} ^{N})\ast \), \(\sigma _{\varepsilon }u_{\varepsilon }^{0}\rightharpoonup v_{0} \) in \(L^{p}(\mathbb{R}^{N})\). Under stronger hypotheses on \(\sigma _{\varepsilon }\) and \(u_{\varepsilon }^{0}\), the author proves that \( u_{\varepsilon }\) converges strongly in \(L^{1}(0;T;L_{\mathrm{loc}}^{2}(\mathbb{R} ^{N}))\) to the solution \(u\) to the transport equation \(\frac{\partial u}{ \partial t}-\frac{\xi _{0}}{\sigma _{0}}\cdot \nabla _{x}u=0\) in \( (0,T)\times \mathbb{R}^{N}\), with the initial condition \(u(0,\cdot )=u^{0}\) in \(\mathbb{R}^{N}\). The author first proves a uniform \(L^{p}\) bound for \( u_{\varepsilon }\): \(\left\Vert u_{\varepsilon }\right\Vert _{L^{p}(\mathbb{R} ^{N})}\leq C\left\Vert u_{\varepsilon }^{0}\right\Vert _{L^{p}(\mathbb{R} ^{N})}\), if \(u_{\varepsilon }^{0}\in L^{p}(\mathbb{R}^{N})\) with \(p\in \lbrack 1,\infty )\). Then he proves that \(W_{\varepsilon }\) and \(W\) are \( C^{1}\)-diffeomorphisms on \(\mathbb{R}^{N}\). The main step of the proof consists to introduce a weak formulation of the transport equation and to pass to the limit in the different terms using uniqueness results by \textit{R. J. DiPerna} and \textit{P. L. Lions} in [Invent. Math. 98, No. 3, 511--547 (1989; Zbl 0696.34049)] and the div-curl Lemma by \\textit{F. Murat} [Ann. Sc. Norm. Super. Pisa, Cl. Sci., IV. Ser. 5, 489--507 (1978; Zbl 0399.46022)], among other tools. In the last part of his paper, the author presents three examples of such situations.