Class AbstractLeastSquaresOptimizer
- All Implemented Interfaces:
DifferentiableMultivariateVectorialOptimizer
- Direct Known Subclasses:
GaussNewtonOptimizer
,LevenbergMarquardtOptimizer
This base class handles the boilerplate methods associated to thresholds settings, jacobian and error estimation.
- Since:
- 1.2
- Version:
- $Revision: 1073158 $ $Date: 2011-02-21 22:46:52 +0100 (lun. 21 févr. 2011) $
-
Field Summary
FieldsModifier and TypeFieldDescriptionprotected VectorialConvergenceChecker
Convergence checker.protected int
Number of columns of the jacobian matrix.protected double
Cost value (square root of the sum of the residuals).static final int
Default maximal number of iterations allowed.protected double[][]
Jacobian matrix.protected double[]
Current objective function value.protected double[]
Current point.protected double[]
Current residuals.protected double[]
Weight for the least squares cost computation.protected int
Number of rows of the jacobian matrix.protected double[]
Target value for the objective functions at optimum.protected double[][]
Weighted Jacobianprotected double[]
Weighted residuals -
Constructor Summary
ConstructorsModifierConstructorDescriptionprotected
Simple constructor with default settings. -
Method Summary
Modifier and TypeMethodDescriptionprotected abstract VectorialPointValuePair
Perform the bulk of optimization algorithm.double
Get a Chi-Square-like value assuming the N residuals follow N distinct normal distributions centered on 0 and whose variances are the reciprocal of the weights.Get the convergence checker.double[][]
Get the covariance matrix of optimized parameters.int
Get the number of evaluations of the objective function.int
Get the number of iterations realized by the algorithm.int
Get the number of evaluations of the objective function jacobian .int
Get the maximal number of functions evaluations.int
Get the maximal number of iterations of the algorithm.double
getRMS()
Get the Root Mean Square value.double[]
Guess the errors in optimized parameters.protected void
Increment the iterations counter by 1.optimize
(DifferentiableMultivariateVectorialFunction f, double[] target, double[] weights, double[] startPoint) Optimizes an objective function.void
setConvergenceChecker
(VectorialConvergenceChecker convergenceChecker) Set the convergence checker.void
setMaxEvaluations
(int maxEvaluations) Set the maximal number of functions evaluations.void
setMaxIterations
(int maxIterations) Set the maximal number of iterations of the algorithm.protected void
Update the jacobian matrix.protected void
Update the residuals array and cost function value.
-
Field Details
-
DEFAULT_MAX_ITERATIONS
public static final int DEFAULT_MAX_ITERATIONSDefault maximal number of iterations allowed.- See Also:
-
checker
Convergence checker. -
jacobian
protected double[][] jacobianJacobian matrix.This matrix is in canonical form just after the calls to
updateJacobian()
, but may be modified by the solver in the derived class (theLevenberg-Marquardt optimizer
does this). -
cols
protected int colsNumber of columns of the jacobian matrix. -
rows
protected int rowsNumber of rows of the jacobian matrix. -
targetValues
protected double[] targetValuesTarget value for the objective functions at optimum.- Since:
- 2.1
-
residualsWeights
protected double[] residualsWeightsWeight for the least squares cost computation.- Since:
- 2.1
-
point
protected double[] pointCurrent point. -
objective
protected double[] objectiveCurrent objective function value. -
residuals
protected double[] residualsCurrent residuals. -
wjacobian
protected double[][] wjacobianWeighted Jacobian -
wresiduals
protected double[] wresidualsWeighted residuals -
cost
protected double costCost value (square root of the sum of the residuals).
-
-
Constructor Details
-
AbstractLeastSquaresOptimizer
protected AbstractLeastSquaresOptimizer()Simple constructor with default settings.The convergence check is set to a
SimpleVectorialValueChecker
and the maximal number of evaluation is set to its default value.
-
-
Method Details
-
setMaxIterations
public void setMaxIterations(int maxIterations) Set the maximal number of iterations of the algorithm.- Specified by:
setMaxIterations
in interfaceDifferentiableMultivariateVectorialOptimizer
- Parameters:
maxIterations
- maximal number of function calls .
-
getMaxIterations
public int getMaxIterations()Get the maximal number of iterations of the algorithm.- Specified by:
getMaxIterations
in interfaceDifferentiableMultivariateVectorialOptimizer
- Returns:
- maximal number of iterations
-
getIterations
public int getIterations()Get the number of iterations realized by the algorithm.- Specified by:
getIterations
in interfaceDifferentiableMultivariateVectorialOptimizer
- Returns:
- number of iterations
-
setMaxEvaluations
public void setMaxEvaluations(int maxEvaluations) Set the maximal number of functions evaluations.- Specified by:
setMaxEvaluations
in interfaceDifferentiableMultivariateVectorialOptimizer
- Parameters:
maxEvaluations
- maximal number of function evaluations
-
getMaxEvaluations
public int getMaxEvaluations()Get the maximal number of functions evaluations.- Specified by:
getMaxEvaluations
in interfaceDifferentiableMultivariateVectorialOptimizer
- Returns:
- maximal number of functions evaluations
-
getEvaluations
public int getEvaluations()Get the number of evaluations of the objective function.The number of evaluation correspond to the last call to the
optimize
method. It is 0 if the method has not been called yet.- Specified by:
getEvaluations
in interfaceDifferentiableMultivariateVectorialOptimizer
- Returns:
- number of evaluations of the objective function
-
getJacobianEvaluations
public int getJacobianEvaluations()Get the number of evaluations of the objective function jacobian .The number of evaluation correspond to the last call to the
optimize
method. It is 0 if the method has not been called yet.- Specified by:
getJacobianEvaluations
in interfaceDifferentiableMultivariateVectorialOptimizer
- Returns:
- number of evaluations of the objective function jacobian
-
setConvergenceChecker
Set the convergence checker.- Specified by:
setConvergenceChecker
in interfaceDifferentiableMultivariateVectorialOptimizer
- Parameters:
convergenceChecker
- object to use to check for convergence
-
getConvergenceChecker
Get the convergence checker.- Specified by:
getConvergenceChecker
in interfaceDifferentiableMultivariateVectorialOptimizer
- Returns:
- object used to check for convergence
-
incrementIterationsCounter
Increment the iterations counter by 1.- Throws:
OptimizationException
- if the maximal number of iterations is exceeded
-
updateJacobian
Update the jacobian matrix.- Throws:
FunctionEvaluationException
- if the function jacobian cannot be evaluated or its dimension doesn't match problem dimension
-
updateResidualsAndCost
Update the residuals array and cost function value.- Throws:
FunctionEvaluationException
- if the function cannot be evaluated or its dimension doesn't match problem dimension or maximal number of of evaluations is exceeded
-
getRMS
public double getRMS()Get the Root Mean Square value. Get the Root Mean Square value, i.e. the root of the arithmetic mean of the square of all weighted residuals. This is related to the criterion that is minimized by the optimizer as follows: if c if the criterion, and n is the number of measurements, then the RMS is sqrt (c/n).- Returns:
- RMS value
-
getChiSquare
public double getChiSquare()Get a Chi-Square-like value assuming the N residuals follow N distinct normal distributions centered on 0 and whose variances are the reciprocal of the weights.- Returns:
- chi-square value
-
getCovariances
Get the covariance matrix of optimized parameters.- Returns:
- covariance matrix
- Throws:
FunctionEvaluationException
- if the function jacobian cannot be evaluatedOptimizationException
- if the covariance matrix cannot be computed (singular problem)
-
guessParametersErrors
Guess the errors in optimized parameters.Guessing is covariance-based, it only gives rough order of magnitude.
- Returns:
- errors in optimized parameters
- Throws:
FunctionEvaluationException
- if the function jacobian cannot b evaluatedOptimizationException
- if the covariances matrix cannot be computed or the number of degrees of freedom is not positive (number of measurements lesser or equal to number of parameters)
-
optimize
public VectorialPointValuePair optimize(DifferentiableMultivariateVectorialFunction f, double[] target, double[] weights, double[] startPoint) throws FunctionEvaluationException, OptimizationException, IllegalArgumentException Optimizes an objective function.Optimization is considered to be a weighted least-squares minimization. The cost function to be minimized is ∑weighti(objectivei-targeti)2
- Specified by:
optimize
in interfaceDifferentiableMultivariateVectorialOptimizer
- Parameters:
f
- objective functiontarget
- target value for the objective functions at optimumweights
- weight for the least squares cost computationstartPoint
- the start point for optimization- Returns:
- the point/value pair giving the optimal value for objective function
- Throws:
FunctionEvaluationException
- if the objective function throws one during the searchOptimizationException
- if the algorithm failed to convergeIllegalArgumentException
- if the start point dimension is wrong
-
doOptimize
protected abstract VectorialPointValuePair doOptimize() throws FunctionEvaluationException, OptimizationException, IllegalArgumentExceptionPerform the bulk of optimization algorithm.- Returns:
- the point/value pair giving the optimal value for objective function
- Throws:
FunctionEvaluationException
- if the objective function throws one during the searchOptimizationException
- if the algorithm failed to convergeIllegalArgumentException
- if the start point dimension is wrong
-