dev.nm.solver.multivariate.geneticalgorithm.minimizer.deoptim

Class DEOptim

java.lang.Object

dev.nm.solver.multivariate.geneticalgorithm.minimizer.simplegrid.SimpleGridMinimizer

dev.nm.solver.multivariate.geneticalgorithm.minimizer.deoptim.DEOptim

All Implemented Interfaces:

Minimizer<OptimProblem,IterativeSolution<Vector>>, IterativeMinimizer<OptimProblem>, MultivariateMinimizer<OptimProblem,IterativeSolution<Vector>>, Optimizer<OptimProblem,IterativeSolution<Vector>>

public class DEOptim
extends SimpleGridMinimizer

Differential Evolution (DE) is a global optimization method. It is a genetic algorithm that iteratively tries to improve a candidate solution with regard to a given measure of quality. DE is used for multidimensional real-valued functions but does not use the gradient of the problem being optimized, which means DE does not require for the optimization problem to be differentiable as is required by classic optimization methods such as gradient descent and quasi-newton methods. DE can therefore also be used on optimization problems that are not even continuous, are noisy, change over time, etc. Comparing to other genetic algorithm optimization methods, the breakthrough of DE is the idea of using vector differences for perturbing the vector population.

The R equivalent function is DEoptim in package DEoptim.

See Also:

• Wikipedia: Differential evolution
• Differential Evolution (DE)
• "Kenneth Price, Rainer M. Storn, Jouni A. Lampinen. Differential Evolution: A Practical Approach to Global Optimization. Springer. 2005."

Nested Class Summary

Nested Classes

Modifier and Type	Class and Description
static interface	DEOptim.NewCellFactory This factory constructs a new DEOptimCellFactory for each minimization problem.
protected class	DEOptim.Solution This is the solution to a minimization problem using DEOptim.

Nested classes/interfaces inherited from class dev.nm.solver.multivariate.geneticalgorithm.minimizer.simplegrid.SimpleGridMinimizer

SimpleGridMinimizer.NewCellFactoryCtor

Field Summary

Fields inherited from class dev.nm.solver.multivariate.geneticalgorithm.minimizer.simplegrid.SimpleGridMinimizer

DEFAULT_STABLE_ITERATION_COUNT, epsilon, factoryCtor, maxIterations, nStableIterations, uniform

Constructor Summary

Constructors

Constructor and Description
DEOptim(DEOptim.NewCellFactory factoryCtor, RandomLongGenerator uniform, double epsilon, int maxIterations, int nStableIterations) Construct a DEOptim to solve unconstrained minimization problems.
DEOptim(double Cr, double F, double epsilon, int maxIterations, int nStableIterations) Construct a DEOptim to solve unconstrained minimization problems.
DEOptim(double Cr, double F, RandomLongGenerator uniform, double epsilon, int maxIterations, int nStableIterations) Construct a DEOptim to solve unconstrained minimization problems.

Method Summary

Modifier and Type	Method and Description
IterativeSolution<Vector>	solve(OptimProblem problem) Solve an optimization problem, e.g., OptimProblem.

Methods inherited from class java.lang.Object

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

Constructor Detail

DEOptim

public DEOptim(DEOptim.NewCellFactory factoryCtor,
               RandomLongGenerator uniform,
               double epsilon,
               int maxIterations,
               int nStableIterations)

Construct a DEOptim to solve unconstrained minimization problems.

Parameters:

factoryCtor - a factory that constructs a new instance of SimpleCellFactory for each problem

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.

DEOptim

public DEOptim(double Cr,
               double F,
               RandomLongGenerator uniform,
               double epsilon,
               int maxIterations,
               int nStableIterations)

Construct a DEOptim to solve unconstrained minimization problems.

Parameters:

Cr - the crossover probability

F - the scaling factor

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.

See Also:

"Kenneth Price, Rainer M. Storn, Jouni A. Lampinen, "Section 2.1.8 Notation," Differential Evolution: A Practical Approach to Global Optimization, Springer, 2005."

DEOptim

public DEOptim(double Cr,
               double F,
               double epsilon,
               int maxIterations,
               int nStableIterations)

Construct a DEOptim to solve unconstrained minimization problems.

Parameters:

Cr - the crossover probability

F - the scaling factor

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.

See Also:

"Kenneth Price, Rainer M. Storn, Jouni A. Lampinen, "Section 2.1.8 Notation," Differential Evolution: A Practical Approach to Global Optimization, Springer, 2005."

Method Detail

solve

public IterativeSolution<Vector> solve(OptimProblem problem)
                                throws Exception

Description copied from interface: Optimizer

Solve an optimization problem, e.g., OptimProblem.

Specified by:

solve in interface Optimizer<OptimProblem,IterativeSolution<Vector>>

Overrides:

solve in class SimpleGridMinimizer

Parameters:

problem - an optimization problem

Returns:

a solution to the optimization problem

Throws:

Exception - when there is an error solving the problem