base namespace for this project. More...

Namespaces
	configuration
	Contains the variables which store information about the configuration for the particle swarm optimization algorithm (PSO).

	parse
	Contains a large set of functions for parsing the configuration file.

	visualization
	This namespace contains functionality to visualize the particle swarm.

Classes
class	AbsOperation
	This operation calculates the absolute value of the supplied value. More...

class	AddOperation
	This pair combination operation reduces two values by addition. More...

class	AdjacencyListNeighborhood
	This class is an abstract class, which implements a complete neighborhood topology through a neighborhood relationship specified by adjacency lists. More...

class	ArccosOperation
	This operation calculates the arccosine function of the supplied value. More...

class	ArcsinOperation
	This operation calculates the arcsine function of the supplied value. More...

class	ArctanOperation
	This operation calculates the arctangent function of the supplied value. More...

class	ArithmeticAverageReduceOperation
	This vector reduce operation reduces a vector by calculation of its arithmetic average. More...

class	BoundHandling
	This class is an abstract class. It supplies the interface for bound handling strategies of the particle swarm optimization algorithm. More...

class	BoundHandlingAbsorption
	This class implements the bound handling strategy absorption for the particle swarm optimization. More...

class	BoundHandlingHyperbolic
	This class implements the bound handling strategy hyperbolic for the particle swarm optimization. More...

class	BoundHandlingNearest
	This class implements the bound handling strategy nearest for the particle swarm optimization. More...

class	BoundHandlingNoBounds
	This class implements the bound handling strategy no bounds for the particle swarm optimization. More...

class	BoundHandlingRandom
	This class implements the bound handling strategy random for the particle swarm optimization. More...

class	BoundHandlingRandomForth
	This class implements the bound handling strategy random forth for the particle swarm optimization. More...

class	BoundHandlingReflect
	This class implements the bound handling strategy reflect for the particle swarm optimization. More...

class	BoundHandlingTorus
	This class implements the bound handling strategy torus for the particle swarm optimization. More...

class	CombineFunction
	This class implements a function which combines two functions with an operation. More...

class	CombineSpecificFunction
	This class implements a specific function, which combines two specific functions with an operation. More...

class	CombineSpecificStatisticalEvaluation
	This class implements a specific statistical evaluation, which combines two specific statistical evaluations by a combination operation. More...

class	CombineStatistic
	This class implements a statistic, which combines two statistics by a combination operation. More...

class	ComposedDimensionReduceOperation
	This class implements a reduce operation, which reduces the values of all dimensions to a single value. More...

class	ComposedParticleReduceOperation
	This class implements a reduce operation, which reduces the values of all particles to a single value. More...

class	ComposedStatisticReduceOperation
	This class is an abstract class. Additionally to the statistic reduce operation it stores a vector reduce operation, which determines how specific statistical evaluations should be reduced. More...

class	ConstantEvaluation
	This class is an abstract class. It supplies the interface for generating a constant value. More...

class	ConstantFunction
	This class implements a specific function with an evaluation function that always evaluates to a vector of a predefined constant. More...

class	ConstantSpecificFunction
	This class implements a specific function with an evaluation function that always evaluates to a vector of a predefined constant. More...

class	ConstantSpecificStatisticalEvaluation
	This class implements a specific statistical evaluation, which always evaluates to the same value. More...

class	ConstantStatistic
	This class implements a statistic, which always evaluates to the same value. More...

class	CosOperation
	This operation calculates the cosine function of the supplied value. More...

class	DefaultUpdater
	This class implements the standard movement equations for particle swarm optimization. More...

class	DeltaUpdateCounterEvaluation
	This class implements a specific statistical evaluation, which supplies the number of delta updates if a DeltaUpdater is used. More...

class	DeltaUpdater
	The delta updater implements a specific position and velocity update procedure. More...

class	DeltaUpdaterTCS
	The delta updater implements a specific position and velocity update procedure. More...

class	DiagonalFunction
	This class implements the diagonal function, which represents a function with optimum at the origin and constant second derivative matrix. More...

class	DimensionIndependentUpdater
	This class implements an update procedure, which acts independent to the coordinate axes. More...

class	DistTo1DOptimumOperation
	This vector operation calculates for each dimension the distance to the nearest 1-dimensional local optimum. More...

class	DivideOperation
	This pair combination operation reduces two values by division. More...

class	DoubleConstantEvaluation
	This constant evaluation can calculate a predefined mpf_t value from a double value. More...

class	EConstantEvaluation
	This constant evaluation calculates the value E. More...

class	ExpOperation
	This operation calculates the exponential function of the supplied value. More...

class	Function
	This class is an abstract class. It supplies the interface for objective functions. More...

class	FunctionEvaluationReduceOperation
	This vector reduce operation reduces a vector by evaluation of the given function. More...

class	FunctionReduceOperator
	This class is a function, which reduces a specific function to a function with a vector reduce operation. More...

class	GeometricAverageReduceOperation
	This vector reduce operation reduces a vector by calculation of its geometric average. More...

class	GlobalAttractorEvaluation
	This class implements a specific statistical evaluation, which supplies the global attractor positions of the particles. More...

class	GlobalAttractorUpdatesStatistic
	This class implements a statistic, which evaluates how often each particle updates a global attractor. More...

class	GlobalBest
	This class implements the global best topology for the particle swarm, where all particles are neighbors. More...

class	GlobalBestPositionDistTo1DOptimumStatistic
	This class implements a statistic, which evaluates the distance of the global best found position to the nearest local optimum if only a single dimension is variable. More...

class	GlobalBestPositionFunctionEvaluationStatistic
	This class implements a statistic, which evaluates the objective function at the global best position found so far. More...

class	GlobalBestPositionStatistic
	This class implements a statistic, which displays the position entries of the overall best position found so far. More...

class	GridNeighborhood
	This class implements the grid or von Neumann topology for the particle swarm. More...

class	IdentitySpecificFunction
	This class implements a specific function with an evaluation function that clones the input. More...

class	InclinedPlane
	This class implements the inclined plane function, which represents a function with constant first derivative. More...

class	InclinedPlaneAsym
	This class implements the asymmetric inclined plane function, which represents a function with constant first derivative, but different constant entries for each dimension. More...

class	IncreasingOrderNthObjectReduceOperation
	This vector reduce operation reduces a vector by sorting the vector and returning the value at the respective position of the sorted vector. More...

class	LBestNeighborhood
	This class implements the lbest neighborhood topology for the particle swarm. More...

class	LocalAttractorEvaluation
	This class implements a specific statistical evaluation, which supplies the local attractor positions of the particles. More...

class	LocalAttractorUpdatesStatistic
	This class implements a statistic, which evaluates how often each particle updates its local attractor. More...

class	Log2DoubleOperation
	This operation calculates logarithm with base two of the supplied value. More...

class	Log2Operation
	This operation calculates the logarithm with base two of the supplied value. More...

class	LogEOperation
	This operation calculates the logarithm with base E of the supplied value. More...

class	MaximalValueReduceOperation
	This vector reduce operation reduces a vector by calculation of its maximal entry. More...

class	MaxOperation
	This pair combination operation reduces two values by calculation of the maximum. More...

class	MinimalValueReduceOperation
	This vector reduce operation reduces a vector by calculation of its minimal entry. More...

class	MinOperation
	This pair combination operation reduces two values by calculation of the minimum. More...

class	MinusInfinityConstantEvaluation
	This constant evaluation calculates the value -infinity. More...

class	MonoSphere
	This class represents a function, which performs like the sphere function in the first dimension and is constant in all other dimensions. More...

class	MovedRosenbrock
	This class represents a moved version of the well known test function rosenbrock, where the optimal value is moved to the origin. More...

class	MultiplyOperation
	This pair combination operation reduces two values by multiplication. More...

class	NamedStatistic
	This class implements a statistic, where the name is set manually. More...

class	Neighborhood
	This class is an abstract class. It supplies the interface for neighborhood topologies of the particle swarm. More...

class	Norm1
	This class represents a function which calculates the Manhattan norm of a vector. More...

class	Norm1Plus2PowerMinusK
	This class represents a function, which calculates the p-norm for all p=1+1/2^exponent. More...

class	Norm2
	This class represents a function, which calculates the 2-norm. More...

class	Norm2PowerK
	This class represents a function which calculates the p-norm for all p which are powers of 2. More...

class	Norm4
	This class represents a function, which calculates the 4-norm. More...

class	Norm8
	This class represents a function, which calculates the 8-norm. More...

class	NormInfinity
	This class represents a function which calculates the infinity norm of a vector. More...

class	ObjectiveFunctionEvaluationReduceOperation
	This vector reduce operation reduces a vector by evaluation of the specified objective function. More...

class	OperatedFunction
	This class is a function, which applies some operation to the result of another function. More...

class	OperatedSpecificFunction
	This class implements a specific function, which combines an operation with a specific function. More...

class	OperatedSpecificStatisticalEvaluation
	This class implements a specific statistical evaluation, which operates on the values of another specific statistical evaluation. More...

class	OperatedStatistic
	This class implements a statistic, which operates on the values of another statistic. More...

class	Operation
	This class is an abstract class. It supplies the interface to operate on single mpf_t values. More...

class	OrientationChangeUpdater
	The orientation change updater implements an update procedure, which changes orientation if some directions have little relative movement. More...

class	PairReduceOperation
	This class is an abstract class. It supplies the interface for reducing two mpf_t values to a single value. More...

class	Particle
	The particles for the particle swarm optimization algorithm. More...

class	PiConstantEvaluation
	This constant evaluation calculates the value Pi. More...

class	PlusInfinityConstantEvaluation
	This constant evaluation calculates the value +infinity. More...

class	PositionAndVelocityUpdater
	This class is an abstract class for processing the velocity and position update of each particle. More...

class	PositionEvaluation
	This class implements a specific statistical evaluation, which supplies the positions of the particles. More...

class	PotentialAbsVelocityPlusDistToGlobalAttractor
	This potential calculator can calculate the scaled absolute value of the velocity plus the distance to the global attractor. More...

class	PotentialCalculator
	This class is an abstract class. It supplies the interface for calculation of some statistical data. More...

class	PotentialFunctionDifference
	This potential calculator can calculate the function difference statistic. More...

class	PotentialSqrtAbsVelocityPlusSqrtDistToGlobalAttractor
	This potential calculator can calculate the scaled square root of the absolute velocity plus the square root of the distance to the global attractor. More...

class	PowOperation
	This operation calculates the power of the supplied value to a given exponent. More...

class	PrecisionStatistic
	This class implements a statistic, which evaluates how the precision of the evaluations evolve. More...

class	ProductReduceOperation
	This vector reduce operation reduces a vector by calculation of its product. More...

class	ProgramVersion
	This class stores a program version. More...

class	RandomPositiveDefiniteSecondDerivative
	This class represents a function with constant random positive (semi) definite derivative. More...

class	Rastrigin
	This class represents the well known test function rastrigin. More...

class	RingNeighborhood
	This class implements the ring neighborhood topology for the particle swarm. More...

class	Rosenbrock
	This class represents the well known test function rosenbrock. More...

class	ScaledHadamardRotatedSphere
	This class represents a scaled and rotated version of the well known sphere function. More...

class	ScaledHadamardRotatedSphereFixed
	This class represents a scaled and rotated version of the well known sphere function with fixed scale 10⁶. More...

class	ScaledSphere
	This class represents a scaled version of the well known sphere function. More...

class	ScaledSphereFixed
	This class represents a scaled version of the well known sphere function with fixed scale 10⁶. More...

class	ScaledSphereRand
	This class represents a randomly scaled version of the well known sphere function. More...

class	Schwefel1
	This class represents the well known test function schwefel. More...

class	Schwefel2
	This class represents the well known test function schwefel. More...

class	SearchSpaceBound
	This class can store information about where the bounds of the objective function are. More...

class	SingleDifferentDirection
	This class implements a function with a special direction. More...

class	SinOperation
	This operation calculates the sine function of the supplied value. More...

class	SortingFunction
	This class implements a function, which measures the sortedness of the position entries. More...

class	SpecificFunction
	This class is an abstract class. It supplies an interface for operations on vectors. More...

class	SpecificIdReduceOperation
	This vector reduce operation reduces a vector by ignoring all values of the vector except the specified one. More...

class	SpecificStatisticalEvaluation
	This class is an abstract class. It supplies the interface for intermediate statistical evaluations. More...

class	Sphere
	This class represents a function which calculates the 2-norm (which is equivalent to the sphere function). More...

class	SpherePlus
	This class represents the well known test function sphere plus. More...

class	SqrtOperation
	This operation calculates the square root of the supplied value. More...

class	Statistic
	This class is an abstract class. It supplies the interface for statistical evaluations. More...

class	StatisticReduceOperation
	This class is an abstract class. It supplies the interface to reduce specific statistical evaluations to analyzable statistics. More...

class	StatisticReduceOperator
	This class implements a statistic, which reduces a specific statistical evaluation to an analyzable statistic. More...

class	Statistics
	This class stores current information about the particle swarm. More...

class	SubtractOperation
	This pair combination operation reduces two values by subtraction. More...

class	SumReduceOperation
	This vector reduce operation reduces a vector by calculation of its sum. More...

class	TanOperation
	This operation calculates the tangent function of the supplied value. More...

class	TestingFunction
	This class is generated for testing purpose. More...

class	TestingUpdater
	The testing updater implements an experimental position and velocity update procedure. More...

class	TwoCupsFunction
	This class represents a test function with 2^D many optimal positions. More...

class	VectorOperation
	This class is an abstract class. It supplies the interface to operate on vectors of mpf_t values. More...

class	VectorReduceOperation
	This an abstract class. It supplies the interface for reducing a vector of mpf_t values to a single value. More...

class	VelocityAdjustment
	This class is an abstract class for processing the velocity adjustment after the position update of each particle. More...

class	VelocityAdjustmentAdjust
	This class implements the adjust velocity adjustment strategy. More...

class	VelocityAdjustmentDeterministicBack
	This class implements the deterministic back velocity adjustment strategy. More...

class	VelocityAdjustmentNone
	This class implements the none velocity adjustment strategy. More...

class	VelocityAdjustmentRandomBack
	This class implements the random back velocity adjustment strategy. More...

class	VelocityAdjustmentZero
	This class implements the zero velocity adjustment strategy. More...

class	VelocityEvaluation
	This class implements a specific statistical evaluation, which supplies the velocities of the particles. More...

class	WheelNeighborhood
	This class implements the wheel topology for the particle swarm. More...

Enumerations
enum	AbsorptionMarkDimensions { ABSORPTION_MARK_DIMENSIONS_ALL_ON_CHANGE, ABSORPTION_MARK_DIMENSIONS_OUTSIDE_BOUNDS, ABSORPTION_MARK_DIMENSIONS_ON_BOUNDARY }
	This enumeration describes, which dimensions are marked as modified, if absorption bound handling is active. This determines, for which dimensions the velocity adjustment is used. More...

enum	RandomForthMarkDimensions { RANDOM_FORTH_MARK_DIMENSIONS_ALL_ON_CHANGE, RANDOM_FORTH_MARK_DIMENSIONS_OUTSIDE_BOUNDS, RANDOM_FORTH_MARK_DIMENSIONS_ON_BOUNDARY }
	This enumeration describes, which dimensions are marked as modified, if random forth bound handling is active. This determines, for which dimensions the velocity adjustment is used. More...

enum	SingleDifferentDirectionMode { SINGLE_DIFFERENT_DIRECTION_MODE_FIRST, SINGLE_DIFFERENT_DIRECTION_MODE_DIAGONAL, SINGLE_DIFFERENT_DIRECTION_MODE_RANDOM }
	This enumeration determines the specific behavior of the single different direction function. More...

Functions
void	__AssertCondition (const char argument, const char file, const char *function_name, int line, std::string message)
	Prints the given information to stderr and let the program terminate with exit code 1. More...

bool	AllowedToRun ()
	If run check is activated then This method checks whether the program is currently allowed to run. More...

Statistics *	DoPso (Statistics *statistics)
	This method proceeds the activities of the PSO algorithm. More...

Statistics *	InitAndDoPso ()
	This method initializes the particles and starts the swarm. More...

bool	operator< (const ProgramVersion &a, const ProgramVersion &b)
	Comparison operator "less" between program versions. More...

bool	operator<= (const ProgramVersion &a, const ProgramVersion &b)
	Comparison operator "less or equal" between program versions. More...

bool	operator== (const ProgramVersion &a, const ProgramVersion &b)
	Comparison operator "equal" between program versions. More...

bool	operator> (const ProgramVersion &a, const ProgramVersion &b)
	Comparison operator "greater" between program versions. More...

bool	operator>= (const ProgramVersion &a, const ProgramVersion &b)
	Comparison operator "greater or equal" between program versions. More...

Statistics *	RestoreAndDoPso ()
	Loads information from the backup file and restarts the particle swarm. More...

void	Shutdown ()
	Stores all data and generates a shutdown file to ease restarting of this particle swarm execution. More...

int	StartTasks (char *argv[])
	Restarts all particle swarms which are correctly terminated in the folder specified in argument 2 of the supplied char * array. More...

void	WriteCurrentState (std::string filename)
	Writes a backup file with the specified file name and stores there all needed data to restart the particle swarm. More...

Variables
const ProgramVersion	PSO_PROGRAM_VERSION
	Describes the current version of this PSO program. More...

Detailed Description

base namespace for this project.

Enumeration Type Documentation

enum highprecisionpso::AbsorptionMarkDimensions

This enumeration describes, which dimensions are marked as modified, if absorption bound handling is active. This determines, for which dimensions the velocity adjustment is used.

Enumerator
ABSORPTION_MARK_DIMENSIONS_ALL_ON_CHANGE	If the position of at least one dimension is situated outside of the search space bounds, the velocity adjustment is applied on all dimensions.
ABSORPTION_MARK_DIMENSIONS_OUTSIDE_BOUNDS	The velocity adjustment is applied only in dimensions, where the position is situated outside of the search space bounds.
ABSORPTION_MARK_DIMENSIONS_ON_BOUNDARY	The velocity adjustment is applied only in dimensions, where the new position is situated exactly on the search space bounds.

enum highprecisionpso::RandomForthMarkDimensions

This enumeration describes, which dimensions are marked as modified, if random forth bound handling is active. This determines, for which dimensions the velocity adjustment is used.

Enumerator
RANDOM_FORTH_MARK_DIMENSIONS_ALL_ON_CHANGE	If the position of at least one dimension is situated outside of the search space bounds, the velocity adjustment is applied on all dimensions.
RANDOM_FORTH_MARK_DIMENSIONS_OUTSIDE_BOUNDS	The velocity adjustment is applied only in dimensions, where the position is situated outside of the search space bounds.
RANDOM_FORTH_MARK_DIMENSIONS_ON_BOUNDARY	The velocity adjustment is applied only in dimensions, where the new position is situated exactly on the search space bounds. (The event that the new position is situated exactly on the search space bounds is very unlikely).

enum highprecisionpso::SingleDifferentDirectionMode

This enumeration determines the specific behavior of the single different direction function.

Enumerator
SINGLE_DIFFERENT_DIRECTION_MODE_FIRST	In this mode the special direction is the direction of the coordinate axis belonging to the first dimension. The special direction vector is then (1,0,...,0)
SINGLE_DIFFERENT_DIRECTION_MODE_DIAGONAL	In this mode the special direction is diagonal according to all dimensions. All entries of the direction vector are 1. The special direction vector is then (1,1,...,1)
SINGLE_DIFFERENT_DIRECTION_MODE_RANDOM	In this mode the special direction is randomly chosen according to the random number generator for this function. The special direction vector is then (r₀,r₁,...,r_D-1), where each r_i is randomly distributed similar to a gaussian distribution with expectation 0 and variance 1.

Function Documentation

void highprecisionpso::__AssertCondition	(	const char *	argument,
		const char *	file,
		const char *	function_name,
		int	line,
		std::string	message
	)

Prints the given information to stderr and let the program terminate with exit code 1.

Parameters

argument	The argument of the checked condition.
file	The file where the error occurs.
function_name	The function name where the error occurs.
line	The line where the error occurs
message	A further information message, which gives some additional information why the error occurs.

bool highprecisionpso::AllowedToRun ( )

If run check is activated then This method checks whether the program is currently allowed to run.

Returns: The truth value whether the program is allowed to run.

Statistics* highprecisionpso::DoPso ( Statistics * statistics )

This method proceeds the activities of the PSO algorithm.

Here the PSO iterations are controlled. At each iteration the update of each position of the particles is started. The actual update of the global attractor is initialized here. Statistical evaluations, the swarm debugging, backups and the check whether the program is allowed to run are started here.

Parameters

statistics The statistics where data about the current PSO execution are stored.

Returns: The final statistic.

Statistics* highprecisionpso::InitAndDoPso ( )

This method initializes the particles and starts the swarm.

Returns: The final statistic.

bool highprecisionpso::operator<	(	const ProgramVersion &	a,
		const ProgramVersion &	b
	)

Comparison operator "less" between program versions.

Parameters

a	The first program version which should be compared.
b	The second program version which should be compared.

Returns: The result of the comparison.