com.numericalmethod.suanshu.optimization.geneticalgorithm.minimizer.local

Class GlobalSearchByLocalMinimizer

java.lang.Object

com.numericalmethod.suanshu.optimization.geneticalgorithm.minimizer.simplegrid.SimpleGridMinimizer

com.numericalmethod.suanshu.optimization.geneticalgorithm.minimizer.local.GlobalSearchByLocalMinimizer

All Implemented Interfaces:

Minimizer<OptimProblem,IterativeMinimizer<Vector>>, Optimizer<OptimProblem,IterativeMinimizer<Vector>>, java.io.Serializable

public class GlobalSearchByLocalMinimizer
extends SimpleGridMinimizer

This minimizer is a global optimization method. It puts a mesh over the feasible region and then locally searches (optimizes) the neighborhood around each mesh point. The algorithm tries to escape the local minimums by crossing over other local minimums using a genetic algorithm. The local search (mutation) is performed by a typical local minimization method, such as Nelder-Mead, BFGS.

See Also:

Serialized Form

Nested Class Summary

Nested classes/interfaces inherited from class com.numericalmethod.suanshu.optimization.geneticalgorithm.minimizer.simplegrid.SimpleGridMinimizer

SimpleGridMinimizer.NewCellFactoryCtor, SimpleGridMinimizer.Solution

Field Summary

Fields inherited from class com.numericalmethod.suanshu.optimization.geneticalgorithm.minimizer.simplegrid.SimpleGridMinimizer

epsilon, factoryCtor, maxIterations, nStableIterations, parallel, uniform

Constructor Summary

Constructors

Constructor and Description
GlobalSearchByLocalMinimizer(boolean parallel) Construct a GlobalSearchByLocalMinimizer to solve unconstrained minimization problems.
GlobalSearchByLocalMinimizer(boolean parallel, RandomLongGenerator uniform, double epsilon, int maxIterations) Construct a GlobalSearchByLocalMinimizer to solve unconstrained minimization problems.
GlobalSearchByLocalMinimizer(LocalSearchCellFactory.MinimizerFactory factory, boolean parallel, RandomLongGenerator uniform, double epsilon, int maxIterations, int nStableIterations) Construct a GlobalSearchByLocalMinimizer to solve unconstrained minimization problems.

Method Summary

Methods inherited from class com.numericalmethod.suanshu.optimization.geneticalgorithm.minimizer.simplegrid.SimpleGridMinimizer

solve

Methods inherited from class java.lang.Object

clone, equals, finalize, getClass, hashCode, notify, notifyAll, toString, wait, wait, wait

Constructor Detail

GlobalSearchByLocalMinimizer

public GlobalSearchByLocalMinimizer(LocalSearchCellFactory.MinimizerFactory factory,
                                    boolean parallel,
                                    RandomLongGenerator uniform,
                                    double epsilon,
                                    int maxIterations,
                                    int nStableIterations)

Construct a GlobalSearchByLocalMinimizer to solve unconstrained minimization problems.

Parameters:

factory - a factory that constructs a new instance of Minimizer for each problem

parallel - true if the algorithm is to run in parallel (multi-core)

uniform - a uniform random number generator

epsilon - a precision parameter: when a number |x| ≤ ε, it is considered 0

maxIterations - the maximum number of iterations

nStableIterations - The solution is considered converged if the minimum does not change over this many iterations.

GlobalSearchByLocalMinimizer

public GlobalSearchByLocalMinimizer(boolean parallel,
                                    RandomLongGenerator uniform,
                                    double epsilon,
                                    int maxIterations)

Construct a GlobalSearchByLocalMinimizer to solve unconstrained minimization problems.

Parameters:

parallel - true if the algorithm is to run in parallel (multi-core)

uniform - a uniform random number generator

epsilon - a precision parameter: when a number |x| ≤ ε, it is considered 0

maxIterations - the maximum number of iterations

GlobalSearchByLocalMinimizer

public GlobalSearchByLocalMinimizer(boolean parallel)

Construct a GlobalSearchByLocalMinimizer to solve unconstrained minimization problems.

Parameters:

parallel - true if the algorithm is to run in parallel (multi-core)