![\[ A^{\prime} A x = A^{\prime} b\]](form_0.png)
for solving the linear least squares problem
![\[ \min_{x} \vert A x - b \vert^2 \]](form_1.png)
for solving the linear least squares problem
as an RVL::Functional | RVLUmin::BacktrackingLineSearchAlg< Scalar > | Factory class for BacktrackingLineSearchAlgBase implementation of backtracking line search |
| RVLUmin::BacktrackingLineSearchAlgBase< Scalar > | Does a backtracking line search starting from a prescribed step, passed as argument firststep to the constructor |
| RVLUmin::CGAlg< Scalar > | Implementation of a CG algorithm |
| RVLUmin::CGException | Exception subtype - thrown when needed |
| RVLUmin::CGNEAlg< Scalar > | Conjugate gradient algorithm - efficient implementation for normal equations
for solving the linear least squares problem
|
| RVLUmin::CGNEPolicy< Scalar > | Policy class for creation of CGNEAlg in trust region solver and any other algorithm needing a least squares solver component - build method creates CGNEAlg with these attributes: |
| RVLUmin::CGNEPolicyData< Scalar > | Data class for CGNE policy |
| RVLUmin::CGNEStep< Scalar > | Single step of conjugate gradient iteration for the normal equations |
| RVLUmin::CGStep< Scalar > | Single iteration of the Conjugate Gradient method for solution of SPD linear systems |
| RVLUmin::ChebAlg< Scalar > | Chebyshev polynomial algorithm - efficient implementation for normal equations
for solving the linear least squares problem
|
| RVLUmin::ChebPolicy< Scalar > | |
| RVLUmin::ChebPolicyData< Scalar > | Policy class for creation of ChebAlg - build method creates ChebAlg with these attributes: |
| RVLUmin::ChebStep< Scalar > | Single step of Chebyshev iteration for the normal equations |
| RVLUmin::LBFGSBT< Scalar > | Limited memory Broyden-Fletcher-Goldfarb-Shanno (LBFGS) quasi-Newton optimization with geometric backtracking line search globalization |
| RVLUmin::LBFGSDir< Scalar > | This algorithm performs a quasi-newton method for minimizing a continuous function |
| RVLUmin::LBFGSOp< Scalar > | LMBFGSOp implements the limited memory BFGS approximation to the inverse Hessian of a twice-differentiable function |
| RVLUmin::LineSearchAlg< Scalar > | Abstract handle class template for line searches |
| RVLUmin::LineSearchAlgBase< Scalar > | Base class for line search algorithms |
| RVLUmin::LSQRAlg< Scalar > | This is Algorithm LSQR as stated in Paige and Saunders, ACM TOMS vol |
| RVLUmin::LSQRPolicy< Scalar > | Policy class for creation of LSQRAlg in trust region solver and any other algorithm needing a least squares solver component - build method creates LSQRAlg with these attributes: |
| RVLUmin::LSQRPolicyData< Scalar > | Data class for LSQR policy |
| RVLUmin::LSQRStep< Scalar > | Single step of LSQR iteration for solution of the normal equations, per Paige & Saunders, ACM TOMS v |
| RVLUmin::PCGNEStep< Scalar > | Preconditioned conjugate gradient iteration for the normal equations |
| RVLUmin::PowerMethod< Scalar > | Power method for finding largest singular value of a linear operator |
| RVLUmin::PowerStep< Scalar > | This Algorithm does a single iteration of the Power Method for estimating the largest singular value of a linear operator |
| RVL::Table | |
| RVLUmin::TRGNAlg< Scalar, Policy > | Trust Region iteration |
| RVLUmin::TRGNStep< Scalar, Policy > | Generic trust region (truncated GN) step |
| RVLUmin::UMinDir< Scalar > | Abstract interface for computation of search directions, in support of descent methods |
| RVLUmin::UMinStepLS< Scalar > | Base class for Unconstrained Minimization step algorithms with globalization via line search |
| RVLUmin::VPM< Scalar, LSPolicy, LSPolicyData > | Given a LinOpValOp F and a Vector d in the range of op, implements the function as an RVL::Functional |
1.4.7