Homogenization of elliptic equations with principal part not in divergence form and hamiltonian with quadratic growth (Q3737849)

From MaRDI portal

!

This is the item page for this Wikibase entity, intended for internal use and editing purposes. Please use the normal view instead:

scientific article; zbMATH DE number 3971268
Language Label Description Also known as
default for all languages
No label defined
    English
    Homogenization of elliptic equations with principal part not in divergence form and hamiltonian with quadratic growth
    scientific article; zbMATH DE number 3971268

      Statements

      Homogenization of elliptic equations with principal part not in divergence form and hamiltonian with quadratic growth (English)
      0 references
      0 references
      0 references
      0 references
      1986
      0 references
      homogenization
      0 references
      Hamiltonian
      0 references
      quadratic growth
      0 references
      Hamilton-Jacobi-Bellman equations
      0 references
      stochastic control
      0 references
      0 references
      We consider the following problem: \[ (0_{\epsilon})\quad - a_{ij}(x,x/\epsilon)\partial^ 2u^{\epsilon}/\partial x_ i \partial x_ j-(1/\epsilon)b_ i(x,x/\epsilon)\partial u^{\epsilon}/\partial x_ i=H(x,x/\epsilon),u^{\epsilon},Du^{\epsilon}) \] in \({\mathcal O}\), \(u_{\epsilon}=0\) on \(\Gamma\). Here the coefficients \(a_{ij}\) and \(b_ i\) are smooth, periodic with respect to the second variable, and the matrix \((a_{ij})_{ij}\) is uniformly elliptic. The Hamiltonian H is locally lipschitz continuous with respect to u and Du, and has quadratic growth in Du. The Hamilton- Jacobi-Bellman equations of some stochastic control problems are of this type. Our aim is to pass to the limit in \((0_{\epsilon})\) as \(\epsilon\) tends to zero. We assume the coefficients \(b_ i\) to be centered with respect to the invariant measure of the problem (see the main assumption (3.13)). Then we derive \(L^{\infty}\), \(H^ 1_ 0\) and \(W^{1,p_ 0}(p_ 0>2)\) estimates for the solutions of \((0_{\epsilon})\). We also prove the following corrector's result: \[ Du^{\epsilon}-Du+D\chi (x,x/\epsilon)Du\to 0\quad in\quad (L^ 2({\mathcal O}))^ N\quad strongly. \] This allows one to pass to the limit in \((0_{\epsilon})\) and to obtain: \[ (0_ 0)\quad -q_{ij}(x) \partial^ 2u/\partial x_ i \partial x_ j-r_ i(x) \partial u/\partial x_ i={\mathcal H}(x,u,Du)\quad in\quad {\mathcal O},\quad u=0\quad on\quad \Gamma. \] This problem is of the same type as the initial one. When \((0_{\epsilon})\) is the Hamilton-Jacobi-Bellman equation of a stochastic control problem, then \((0_ 0)\) is also a Hamilton-Jacobi-Bellman equation but corresponding to modified set of controls.
      0 references

      Identifiers

      0 references
      0 references
      0 references
      0 references
      0 references