Class BaseMultiStartMultivariateOptimizer<P>
 Type Parameters:
P
 Type of the point/value pair returned by the optimization algorithm.
 Direct Known Subclasses:
MultiStartMultivariateOptimizer
This class wraps an optimizer in order to use it several times in turn with different starting points (trying to avoid being trapped in a local extremum when looking for a global one). It is not a "user" class.

Field Summary
Fields inherited from class org.hipparchus.optim.BaseOptimizer
evaluations, iterations

Constructor Summary
ConstructorDescriptionBaseMultiStartMultivariateOptimizer
(BaseMultivariateOptimizer<P> optimizer, int starts, RandomVectorGenerator generator) Create a multistart optimizer from a singlestart optimizer. 
Method Summary
Modifier and TypeMethodDescriptionprotected abstract void
clear()
Method that will called in order to clear all stored optima.protected P
Performs the bulk of the optimization algorithm.int
Gets the number of evaluations of the objective function.abstract P[]
Gets all the optima found during the last call tooptimize
.optimize
(OptimizationData... optData) Stores data and performs the optimization.protected abstract void
Method that will be called in order to store each found optimum.Methods inherited from class org.hipparchus.optim.BaseMultivariateOptimizer
getLowerBound, getStartPoint, getUpperBound, parseOptimizationData
Methods inherited from class org.hipparchus.optim.BaseOptimizer
getConvergenceChecker, getIterations, getMaxEvaluations, getMaxIterations, incrementEvaluationCount, incrementIterationCount, optimize

Constructor Details

BaseMultiStartMultivariateOptimizer
public BaseMultiStartMultivariateOptimizer(BaseMultivariateOptimizer<P> optimizer, int starts, RandomVectorGenerator generator) Create a multistart optimizer from a singlestart optimizer.Note that if there are bounds constraints (see
BaseMultivariateOptimizer.getLowerBound()
andBaseMultivariateOptimizer.getUpperBound()
), then a simple rejection algorithm is used at each restart. This implies that the random vector generator should have a good probability to generate vectors in the bounded domain, otherwise the rejection algorithm will hit theBaseOptimizer.getMaxEvaluations()
count without generating a proper restart point. Users must be take great care of the curse of dimensionality. Parameters:
optimizer
 Singlestart optimizer to wrap.starts
 Number of starts to perform. Ifstarts == 1
, theoptimize
will return the same solution as the givenoptimizer
would return.generator
 Random vector generator to use for restarts. Throws:
MathIllegalArgumentException
 ifstarts < 1
.


Method Details

getEvaluations
public int getEvaluations()Gets the number of evaluations of the objective function. The number of evaluations corresponds to the last call to theoptimize
method. It is 0 if the method has not been called yet. Overrides:
getEvaluations
in classBaseOptimizer<P>
 Returns:
 the number of evaluations of the objective function.

getOptima
Gets all the optima found during the last call tooptimize
. The optimizer stores all the optima found during a set of restarts. Theoptimize
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 theoptimize
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 bynull
elements corresponding to the runs that did not converge. This means all elements will benull
if theoptimize
method did throw an exception. This also means that if the first element is notnull
, it is the best point found across all starts.
The behaviour is undefined if this method is called beforeoptimize
; it will likely throwNullPointerException
. Returns:
 an array containing the optima sorted from best to worst.

optimize
Stores data and performs the optimization.The list of parameters is openended so that subclasses can extend it with arguments specific to their concrete implementations.
When the method is called multiple times, instance data is overwritten only when actually present in the list of arguments: when not specified, data set in a previous call is retained (and thus is optional in subsequent calls).
Important note: Subclasses must override
BaseOptimizer.parseOptimizationData(OptimizationData[])
if they need to register their own options; but then, they must also callsuper.parseOptimizationData(optData)
within that method. Overrides:
optimize
in classBaseMultivariateOptimizer<P>
 Parameters:
optData
 Optimization data. In addition to those documented inBaseOptimizer
, this method will register the following data: Returns:
 a point/value pair that satisfies the convergence criteria.
 Throws:
MathIllegalStateException
 ifoptData
does not contain an instance ofMaxEval
orInitialGuess
.

doOptimize
Performs the bulk of the optimization algorithm. Specified by:
doOptimize
in classBaseOptimizer<P>
 Returns:
 the point/value pair giving the optimal value of the objective function.

store
Method that will be called in order to store each found optimum. Parameters:
optimum
 Result of an optimization run.

clear
protected abstract void clear()Method that will called in order to clear all stored optima.
