Checks whether an approximate Hessian can be computed for this problem. function candoit = canGetApproxHessian(problem) Returns true if an approximate Hessian of the cost function is provided in the given problem description, false otherwise. If a Hessian is defined but no approximate Hessian is defined explicitly, returns false. Even if this returns false, calls to getApproxHessian may succeed, as they will be redirected to getHessianFD. The latter simply requires availability of gradients in problem, and vector transports in problem.M. See also: canGetHessian getHessianFD

- preconhessiansolve Preconditioner based on the inverse Hessian, by solving linear systems.
- trustregions Riemannian trust-regions solver for optimization on manifolds.

0001 function candoit = canGetApproxHessian(problem) 0002 % Checks whether an approximate Hessian can be computed for this problem. 0003 % 0004 % function candoit = canGetApproxHessian(problem) 0005 % 0006 % Returns true if an approximate Hessian of the cost function is provided 0007 % in the given problem description, false otherwise. 0008 % If a Hessian is defined but no approximate Hessian is defined explicitly, 0009 % returns false. 0010 % 0011 % Even if this returns false, calls to getApproxHessian may succeed, as 0012 % they will be redirected to getHessianFD. The latter simply requires 0013 % availability of gradients in problem, and vector transports in problem.M. 0014 % 0015 % See also: canGetHessian getHessianFD 0016 0017 % This file is part of Manopt: www.manopt.org. 0018 % Original author: Nicolas Boumal, April 8, 2015. 0019 % Contributors: 0020 % Change log: 0021 0022 candoit = isfield(problem, 'approxhess'); 0023 0024 end

Generated on Sat 12-Nov-2016 14:11:22 by