Package parsimony :: Package functions :: Module penalties :: Class SufficientDescentCondition

Class SufficientDescentCondition

           object --+    
                    |    
  properties.Function --+
                        |
           object --+   |
                    |   |
properties.Constraint --+
                        |
                       SufficientDescentCondition

Nested Classes

[hide private]

Inherited from properties.Function: __metaclass__

Instance Methods

[hide private]

__init__(self, function, p, c=0.0001)
The sufficient condition

source code

f(self, x, a)
Function value.

source code

feasible(self, xa)
Feasibility of the constraint at point x with step a.

source code

Inherited from properties.Function: get_params, reset, set_params

Inherited from object: __delattr__, __format__, __getattribute__, __hash__, __new__, __reduce__, __reduce_ex__, __repr__, __setattr__, __sizeof__, __str__, __subclasshook__

Class Variables

[hide private]

__abstractmethods__ = frozenset([])

Inherited from properties.Function (private): _abc_cache, _abc_negative_cache, _abc_negative_cache_version, _abc_registry

Properties

[hide private]

Inherited from object: __class__

Method Details

[hide private]

init(self, function, p, c=0.0001)
(Constructor)

source code

The sufficient condition

    f(x + a * p) <= f(x) + c * a * grad(f(x))'p

for descent. This condition is sometimes called the Armijo condition.

Parameters
----------
p : Numpy array. The descent direction.

c : Float, 0 < c < 1. A constant for the condition. Should be small.

Overrides: object.__init__

f(self, x, a)

source code

Function value.

Overrides: properties.Function.f: (inherited documentation)

feasible(self, xa)

source code

Feasibility of the constraint at point x with step a.

From the interface "Constraint".

Overrides: properties.Constraint.feasible

Class SufficientDescentCondition

__init__(self, function, p, c=0.0001) (Constructor)

f(self, x, a)

feasible(self, xa)

init(self, function, p, c=0.0001)
(Constructor)