Package minfx :: Module base_classes :: Class Conjugate_gradient

[frames] | no frames]

Class Conjugate_gradient

Instance Methods

__init__(self)
Class containing the non-specific conjugate gradient code.

source code

new_param_func(self)
The new parameter function.

source code

old_cg_conv_test(self)
Convergence tests.

source code

update(self)
Function to update the function value, gradient vector, and Hessian matrix

source code

Method Details

new_param_func(self)

The new parameter function.

Do a line search then calculate xk+1, fk+1, and gk+1.

old_cg_conv_test(self)

Convergence tests.

This is old code implementing the conjugate gradient convergence test given on page 124 of 'Numerical Optimization' by Jorge Nocedal and Stephen J. Wright, 1999, 2nd ed. This function is currently unused.