Stochastic homogenization of certain nonconvex Hamilton-Jacobi equations (Q2420505): Difference between revisions

The author considers the problem of stochastic homogenization for non-convex Hamilton-Jacobi equation. Let \(\check{H}_{i}(p,x,\omega), i=1,\ldots,l\) be quasiconvex Hamiltonians, and let \(\hat{H}_{i}(p,x,\omega), i=1,\ldots,l\) be quasiconcave Hamiltonians. The author considers a non-convex Hamiltonian \(H_{l}(p,x,\omega)\) generated through a min-max formula \(H_{l}:=\max\{\check{H}_{l},\min\{\hat H_{l},\ldots\max\{\check H_2,\min\{\hat H_2,\max\{\check H_1,\hat H_1\}\}\}\ldots\}\}\). It is assumed that \(\check H_1\ge \check H_2\ge\cdots\ge\check H_{l}\), \(\hat H_1\le\hat H_2\le\cdots\le\hat H_{l}\). Under conditions of stationary ergodicity, coercivity, continuity, boundedness, stably pairing and monotonicity on \(\check{H}_{i}\) and \(\hat{H}_{i}\), the following result is proved. For any \(\epsilon>0\) and \(\omega\in \Omega\), let \(u^{\epsilon}(x,t,\omega)\) be the unique viscosity solution of the Hamilton-Jacobi equation \(u^{\epsilon}_{t}+H_{l}(Du^{\epsilon},x/\epsilon,\omega)=0, (x,t)\in \mathbb{R}^{d}\times(0,\infty)\), \(u^{\epsilon}(x,0,\omega)=u_0(x), x\in \mathbb{R}^{d}\). Then there exists an effective Hamiltonian \(\bar H_{l}(p)\in C(\mathbb{R}^{d})\) with \(\lim_{\vert p\vert\to\infty}\bar H_{l}(p)=\infty\), such that for a.e. \(\omega\in \Omega\), \(\lim_{\epsilon\to0}u^{\epsilon}(x,t,\omega)=\bar u(x,t)\) locally uniformly in \(\mathbb{R}^{d}\times(0,\infty)\), where \(\bar u(x,t)\) is the unique viscosity solution of the homogenized Hamilton-Jacobi equation \(\bar u_{t}+\bar H_{l}(D\bar u)=0\), \((x,t)\in \mathbb{R}^{d}\times(0,\infty)\), \(\bar u(x,0)=u_0(x)\), \(x\in \mathbb{R}^{d}\).