Package minfx :: Module steihaug_cg :: Class Steihaug
[hide private]
[frames] | no frames]

Class Steihaug

source code


Instance Methods [hide private]
 
__init__(self, func, dfunc, d2func, args, x0, func_tol, grad_tol, maxiter, epsilon, delta_max, delta0, eta, full_output, print_flag, print_prefix)
Class for Steihaug conjugate-gradient trust region minimisation specific functions.
source code
 
get_pk(self)
The CG-Steihaug algorithm.
source code
 
get_tau(self)
Function to find tau such that p = pj + tau.dj, and ||p|| = delta.
source code
 
new_param_func(self)
Find the CG-Steihaug minimiser.
source code
 
update(self)
Update function.
source code

Inherited from base_classes.Trust_region: trust_region_update

Inherited from newton.Newton: setup_newton, update_newton

Inherited from base_classes.Hessian_mods: cholesky_mod, eigenvalue, gmw, gmw_old, se99, setup_hessian_mod, unmodified_hessian, valid_hessian_mod

Inherited from base_classes.Line_search: backline, line_search_options, mt, no_search, nwi, nww, setup_line_search, valid_line_search

Inherited from base_classes.Min: double_test, func_test, grad_test, hessian_type_and_mod, minimise, setup_conv_tests

Method Details [hide private]

__init__(self, func, dfunc, d2func, args, x0, func_tol, grad_tol, maxiter, epsilon, delta_max, delta0, eta, full_output, print_flag, print_prefix)
(Constructor)

source code 

Class for Steihaug conjugate-gradient trust region minimisation specific functions.

Unless you know what you are doing, you should call the function 'steihaug' rather than using this class.

Overrides: base_classes.Min.__init__

new_param_func(self)

source code 

Find the CG-Steihaug minimiser.

Overrides: newton.Newton.new_param_func

update(self)

source code 

Update function.

Run the trust region update. If this update decides to shift xk+1 to xk, then run the Newton update.

Overrides: base_classes.Min.update