Searched refs:H_k (Results 1 – 2 of 2) sorted by relevance
| /petsc/src/ksp/ksp/utils/lmvm/tests/ |
| H A D | ex1.c | 417 Mat H_k; in TestUpdate() local
424 PetscCall(MatComputeInverseOperator(B, &H_k, PETSC_FALSE)); in TestUpdate()
428 PetscCall(MatAXPY(H_k_exp, -1.0, H_k, SAME_NONZERO_PATTERN)); in TestUpdate()
430 …PetscCall(PetscInfo((PetscObject)H_k, "Forward update error %g, relative error %g\n", (double)err,… in TestUpdate()
431 …PetscCheck(err <= PETSC_SMALL * norm, PetscObjectComm((PetscObject)H_k), PETSC_ERR_PLIB, "Inverse … in TestUpdate()
434 PetscCall(MatDestroy(&H_k)); in TestUpdate()
|
| /petsc/doc/manual/ |
| H A D | tao.md | 891 H_k d_k = -g_k
894 to obtain a step $d_k$, where $H_k$ is the Hessian of the
900 (H_k + \rho_k I) d_k = -g_k
1263 \min_d & \frac{1}{2}d^T H_k d + g_k^T d \\
1268 to obtain a direction $d_k$, where $H_k$ is the Hessian of
1450 $H_k d = -g_k$. The method used to solve the system of equations
1535 H_k d_k = -\nabla f(x_k),
1538 where $H_k$ is the Hessian approximation obtained by using the
1539 BFGS update formula. The inverse of $H_k$ can readily be applied
1719 H_k p_k = -g_k,
[all …]
|