org.apache.commons.math3.optimization
Class BaseMultivariateVectorMultiStartOptimizer<FUNC extends MultivariateVectorFunction>

java.lang.Object
  extended by org.apache.commons.math3.optimization.BaseMultivariateVectorMultiStartOptimizer<FUNC>
Type Parameters:
FUNC - Type of the objective function to be optimized.
All Implemented Interfaces:
BaseMultivariateVectorOptimizer<FUNC>, BaseOptimizer<PointVectorValuePair>
Direct Known Subclasses:
DifferentiableMultivariateVectorMultiStartOptimizer

public class BaseMultivariateVectorMultiStartOptimizer<FUNC extends MultivariateVectorFunction>
extends Object
implements BaseMultivariateVectorOptimizer<FUNC>

Base class for all implementations of a multi-start optimizer. This interface is mainly intended to enforce the internal coherence of Commons-Math. Users of the API are advised to base their code on DifferentiableMultivariateVectorMultiStartOptimizer.

Since:
3.0
Version:
$Id$

Field Summary
private  RandomVectorGenerator generator
          Random generator for multi-start.
private  int maxEvaluations
          Maximal number of evaluations allowed.
private  PointVectorValuePair[] optima
          Found optima.
private  BaseMultivariateVectorOptimizer<FUNC> optimizer
          Underlying classical optimizer.
private  int starts
          Number of starts to go.
private  int totalEvaluations
          Number of evaluations already performed for all starts.
 
Constructor Summary
protected BaseMultivariateVectorMultiStartOptimizer(BaseMultivariateVectorOptimizer<FUNC> optimizer, int starts, RandomVectorGenerator generator)
          Create a multi-start optimizer from a single-start optimizer.
 
Method Summary
 ConvergenceChecker<PointVectorValuePair> getConvergenceChecker()
          Get the convergence checker.
 int getEvaluations()
          Get the number of evaluations of the objective function.
 int getMaxEvaluations()
          Get the maximal number of function evaluations.
 PointVectorValuePair[] getOptima()
          Get all the optima found during the last call to optimize.
 PointVectorValuePair optimize(int maxEval, FUNC f, double[] target, double[] weights, double[] startPoint)
          Optimize an objective function.
private  void sortPairs(double[] target, double[] weights)
          Sort the optima from best to worst, followed by null elements.
 
Methods inherited from class java.lang.Object
clone, equals, finalize, getClass, hashCode, notify, notifyAll, toString, wait, wait, wait
 

Field Detail

optimizer

private final BaseMultivariateVectorOptimizer<FUNC extends MultivariateVectorFunction> optimizer
Underlying classical optimizer.


maxEvaluations

private int maxEvaluations
Maximal number of evaluations allowed.


totalEvaluations

private int totalEvaluations
Number of evaluations already performed for all starts.


starts

private int starts
Number of starts to go.


generator

private RandomVectorGenerator generator
Random generator for multi-start.


optima

private PointVectorValuePair[] optima
Found optima.

Constructor Detail

BaseMultivariateVectorMultiStartOptimizer

protected BaseMultivariateVectorMultiStartOptimizer(BaseMultivariateVectorOptimizer<FUNC> optimizer,
                                                    int starts,
                                                    RandomVectorGenerator generator)
Create a multi-start optimizer from a single-start optimizer.

Parameters:
optimizer - Single-start optimizer to wrap.
starts - Number of starts to perform. If starts == 1, the optimize will return the same solution as optimizer would.
generator - Random vector generator to use for restarts.
Throws:
NullArgumentException - if optimizer or generator is null.
NotStrictlyPositiveException - if starts < 1.
Method Detail

getOptima

public PointVectorValuePair[] getOptima()
Get all the optima found during the last call to optimize. The optimizer stores all the optima found during a set of restarts. The optimize method returns the best point only. This method returns all the points found at the end of each starts, including the best one already returned by the optimize method.
The returned array as one element for each start as specified in the constructor. It is ordered with the results from the runs that did converge first, sorted from best to worst objective value (i.e. in ascending order if minimizing and in descending order if maximizing), followed by and null elements corresponding to the runs that did not converge. This means all elements will be null if the optimize method did throw a ConvergenceException). This also means that if the first element is not null, it is the best point found across all starts.

Returns:
array containing the optima
Throws:
MathIllegalStateException - if optimize has not been called.

getMaxEvaluations

public int getMaxEvaluations()
Get the maximal number of function evaluations.

Specified by:
getMaxEvaluations in interface BaseOptimizer<PointVectorValuePair>
Returns:
the maximal number of function evaluations.

getEvaluations

public int getEvaluations()
Get the number of evaluations of the objective function. The number of evaluations corresponds to the last call to the optimize method. It is 0 if the method has not been called yet.

Specified by:
getEvaluations in interface BaseOptimizer<PointVectorValuePair>
Returns:
the number of evaluations of the objective function.

getConvergenceChecker

public ConvergenceChecker<PointVectorValuePair> getConvergenceChecker()
Get the convergence checker.

Specified by:
getConvergenceChecker in interface BaseOptimizer<PointVectorValuePair>
Returns:
the object used to check for convergence.

optimize

public PointVectorValuePair optimize(int maxEval,
                                     FUNC f,
                                     double[] target,
                                     double[] weights,
                                     double[] startPoint)
Optimize 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 interface BaseMultivariateVectorOptimizer<FUNC extends MultivariateVectorFunction>
Parameters:
maxEval - Maximum number of function evaluations.
f - Objective function.
target - Target value for the objective functions at optimum.
weights - Weights for the least squares cost computation.
startPoint - Start point for optimization.
Returns:
the point/value pair giving the optimal value for objective function.

sortPairs

private void sortPairs(double[] target,
                       double[] weights)
Sort the optima from best to worst, followed by null elements.

Parameters:
target - Target value for the objective functions at optimum.
weights - Weights for the least-squares cost computation.


Copyright (c) 2003-2013 Apache Software Foundation