.. _responses-quasi_hessians-sr1: """ sr1 """ Use the Symmetric Rank 1 update method to compute quasi-Hessians .. toctree:: :hidden: :maxdepth: 1 **Specification** - *Alias:* None - *Arguments:* None **Description** The Symmetric Rank 1 (SR1) update (specified with the keyword ``sr1``) will be used to compute quasi-Hessians. .. math:: B_{k+1} = B_k + \frac{(y_k - B_k s_k)(y_k - B_k s_k)^T}{(y_k - B_k s_k)^T s_k} where :math:`B_k` is the :math:`k^{th}` approximation to the Hessian, :math:`s_k = x_{k+1} - x_k` is the step and :math:`y_k = \nabla f_{k+1} - \nabla f_k` is the corresponding yield in the gradients. *Notes* - Initial scaling of :math:`\frac{y_k^T y_k}{y_k^T s_k} I` is used for :math:`B_0` prior to the first update. - Numerical safeguarding is used to protect against numerically small denominators within the updates. - This safeguarding skips the update if :math:`|(y_k - B_k s_k)^T s_k| < 10^{-6} ||s_k||_2 ||y_k - B_k s_k||_2`