Convergence of the porous media equation to Hele-Shaw (Q5940178)

Let \(\Omega\) be a bounded or an exterior domain in \(\mathbb{R}^n\) and put \(Q= \Omega\times [0,\infty)\), \(Q_T= \Omega\times [0,T]\), \(0< T< \infty\). The connection between the following two problems, the Hele-Shaw problem \((\text{HS}) (u_0,g)\) and porous media equation \((\text{PM})(m,v_0,g)\), is considered: \[ (\text{HS})(u_0, g)\quad\begin{cases} u_t= \Delta v,\quad u\in H(v),\quad &(x,t)\in Q,\\ v|_{\partial\Omega}= g,\quad &t\in (0,\infty),\\ u(x,0)= u_0(x),\quad &x\in \Omega.\end{cases} \] Here \(H(v)\) is a multivalued function defined by \(H(v)= [0,1]\) for \(v=0\) and \(H(v)= 1\) for \(v> 0\). \[ (\text{PM})(m, v_0,g)\quad\begin{cases} (u_m)_t= \Delta(u_m)^m,\quad &(x,t)\in Q,\\ v_m|_{\partial\Omega}= g,\quad & t\in (0,\infty),\\ v_m(x,0)= v_0(x),\quad & x\in \Omega.\end{cases} \] Here \(v_m\) denotes the pressure given by \(v_m= {m\over m-1} (u_m)^{m-1}\). The main theorem asserts that as \(m\) tends to infinity, the profiles of the density and the pressure \((u_m, v_m)\) determined by the unique solution \(u_m\) of \((\text{PM})(m,v_0, g)\) converge to the unique solution \((u,v)\) of \((\text{HS})(u_0, g)\) with \(u_0= \chi_{\{v_0> 0\}}\) (i.e., \(u_0(x)= 0\) if \(v_0(x)= 0\) and \(u_0(x)= 1\) if \(v_0(x)> 0\)) in the strong topology of \(L^1(Q_T)\), provided that \(0\leq v_0\in L^\infty(\Omega)\) and the support of \(v_0\) is compact and that \(0\leq g= g(x)\) can be extended as a function on \(\Omega\) belonging to \(W^{1,2}(\Omega)\cap L^\infty(\Omega)\). Since both equations do not admit classical global solutions in general, the above fact is discussed in a suitable class of weak solutions. For \(0\leq u_0\in L^2(\Omega)\), \(0\leq g= g(x,t)\in L^2(0,\infty; W^{1,2}(\Omega))\), \((u,v)\) is said to be a weak solution of \((\text{HS})(u_0, g)\) iff \(u\in L^2_{\text{loc}}([0,\infty); L^2(\Omega))\), \(v\in L^2_{\text{loc}}([0,\infty); W^{1,2}(\Omega))\), \(\int_{Q_T} (u\phi_t-\nabla v\cdot\nabla\phi) dx dt+ \int_\Omega u_0\phi(x, 0) dx= 0\) holds for all test functions \(\phi\in C^{2,1}(\overline Q_T)\) vanishing on \(\partial\Omega\times [0,T]\) for all \(T> 0\), \(u\in H(v)\) for a.e. \((x,t)\in Q\), and \(v- g\in L^2_{\text{loc}}([0, \infty); W^{1,2}_0(\Omega))\). The uniqueness of weak solution of \((\text{HS})(u_0, g)\) within this class is given. The proof of the main theorem relies on the uniqueness of weak solutions for \((\text{HS})(u_0, g)\) as well as on some a priori estimates for solutions \(u_m\) of \((\text{PM})(m,v_0, g)\) independent of \(m\), which are derived by using the comparison theorem, the energy estimates and the semiconvexity estimate \((\partial(u^m_m)\geq -(u^m_m)/t)\). As a corollary, a comparison theorem for the Hele-Shaw problem and the monotonicity results, i.e., \(u_t\geq 0\), \(v_t\geq 0\) in \({\mathcal D}'(Q)\), are derived for the compactly supported initial data \(0\leq u_0\leq 1\) and boundary data \(g= g(x)\).