Imp
[using it]
 Interface Manipulation Package v4.0 (2008-06-25) a-projector ©copyright Author

imp.estimator
Class Model

java.lang.Object
  extended by imp.estimator.Constraint
      extended by imp.estimator.Model

public abstract class Model
extends Constraint

Defines a local solver of a set of measures.

Nested Class Summary
 class Model.Estimate
          Defines an estimate of the state of the model.
static class Model.Samplings
          Defines a set samplings of the state estimation.
 
Constructor Summary
Model(double step, double[] d_x, double[] d_m)
          Constructs a vectorial model.
 
Method Summary
 Model addMeasure(double[] m)
          Adds a measure and return this.
 Model delMeasure(double[] m)
          Removes a measure and return this.
 Model delMeasures()
          Removes all measures and return this.
 double[][] F_m(double[] m, double[] x)
          Defines the gradients of the model F_m(m, x) = d f(m, x) / d m.
 double[][] F_x(double[] m, double[] x)
          Defines the gradients of the model F_x(m, x) = d f(m, x) / d x.
abstract  double[] f(double[] m, double[] x)
          Defines the model equations f(m, x) = 0.
abstract  int getMeasureNumber()
          Returns the minimal number of generic measures required to estimate a state.
 double[][] getMeasures()
          Gets the measures.
 Model getParent()
          Gets the parent model if any, else return null.
 double[] getState(Model parent, double[] parentState)
          Returns the model initial state using the state of a parent model.
 Model setState(double[] x0)
          Sets the model initial state and return this.
 Model setState(Model parent)
          Sets the model initial state calling the parent model state estimation and return this.
 
Methods inherited from class imp.estimator.Constraint
C, solve
 
Methods inherited from class java.lang.Object
equals, getClass, hashCode, notify, notifyAll, wait, wait, wait
 

Constructor Detail

Model

public Model(double step,
             double[] d_x,
             double[] d_m)
Constructs a vectorial model.

Parameters:
step - Minimal distance between two distinguishable points.
d_x - [optional, default is 1] Precision of the x[] vector.
As double:
If d_x is a double, it defines a uniform precision (i.e. a vector with identical values).
d_m - [optional, default is 1] Precision of the m[] vector.
As double:
If d_m is a double, it defines a uniform precision (i.e. a vector with identical values).
Method Detail

addMeasure

public Model addMeasure(double[] m)
Adds a measure and return this.

delMeasure

public Model delMeasure(double[] m)
Removes a measure and return this.

delMeasures

public Model delMeasures()
Removes all measures and return this.

F_m

public double[][] F_m(double[] m,
                      double[] x)
Defines the gradients of the model F_m(m, x) = d f(m, x) / d m.
The default implementation is a numerical gradient calculation but an analytic evaluation is to be preferred.

F_x

public double[][] F_x(double[] m,
                      double[] x)
Defines the gradients of the model F_x(m, x) = d f(m, x) / d x.
The default implementation is a numerical gradient calculation but an analytic evaluation is to be preferred.

f

public abstract double[] f(double[] m,
                           double[] x)
Defines the model equations f(m, x) = 0.

getMeasureNumber

public abstract int getMeasureNumber()
Returns the minimal number of generic measures required to estimate a state.

getMeasures

public double[][] getMeasures()
Gets the measures.

Returns:
An array of getMeasuresSize() of double[] containing the measures references.

getParent

public Model getParent()
Gets the parent model if any, else return null.

getState

public double[] getState(Model parent,
                         double[] parentState)
Returns the model initial state using the state of a parent model.
This method is used to define how a parent model set this model initial state.

Throws:
IllegalArgumentException - If the initial state definition from the parent model is not defined.

setState

public Model setState(double[] x0)
Sets the model initial state and return this.

setState

public Model setState(Model parent)
Sets the model initial state calling the parent model state estimation and return this.