Package PyDSTool :: Package Toolbox :: Package optimizers :: Package line_search :: Module scaled_line_search :: Class ScaledLineSearch
[hide private]
[frames] | no frames]

Class ScaledLineSearch

source code

object --+
         |
        ScaledLineSearch

A simple line search, takes a point, adds a step and returns it Scales step according to given scales of the parameters and ignores *magnitude* of gradient.

(in early development and experimental only at this point)

Instance Methods [hide private]
 
__init__(self, max_step=1, step_mod=3, **kwargs)
Needs to have :...
source code
 
__call__(self, origin, state, **kwargs)
Returns a good candidate...
source code

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

Properties [hide private]

Inherited from object: __class__

Method Details [hide private]

__init__(self, max_step=1, step_mod=3, **kwargs)
(Constructor)

source code 

Needs to have :
- nothing
Can have :
- max_step: a maximum step control, a scalar or vector to restrict step size
in each direction (default 1)
- step_mod: a factor to divide the step when back-tracking (default 3)
- max_reduce_fac: max_step divided by this is the smallest step that will be tried (default 2000),

Overrides: object.__init__

__call__(self, origin, state, **kwargs)
(Call operator)

source code 

Returns a good candidate
Parameters :
- origin is the origin of the search
- state is the state of the optimizer