For the purpose of parameter evolution EsObject adds a real valued parameter
vector _gene to the class.
The inherited pure virtual methods _mutate() and evaluate() get implemented.
The new method setErrorFunction() is added. To make the class functional derived classes will normally override evaluate()
(quality evaluation) and may override _mutate() (mutation function called by create()).
Methods
|
|
__init__
_mutate
evaluate
setErrFun
str
|
|
__init__
|
__init__ (
self,
Gene=None,
Generation=0,
StepSize=0.1,
)
|
|
_mutate
|
_mutate ( self )
Individual step size will be modified randomly and gene vector
will be variated by adding a random mutation vector with normal
distributed components. The expectation value for the length of the mutation vector is the
applied step size (resulting in a hyper-sphere border distribution).
|
|
evaluate
|
evaluate ( self )
Do the mapping from the genotype _gene to phenotype _phene
assign a quality (fitness) _q Quality function defaults to hyper sphere model with optimum in the Origin.
See also setErrorFunction().
|
|
setErrFun
|
setErrFun ( Class, Argument )
Assign a quality evaluation function (used by evaluate()).
Argument may be either a function or a OptimizationModel instance.
\sa module OptimizationModel
|
|
str
|
str ( self )
|
|