RosenbrockEvaluator Class

Namespace: Phoenix.Optimization.AlgorithmTests.Evaluators

Overview

Implementation of Rosenbrock Fucntion Source: http://www-optima.amp.i.kyoto-u.ac.jp/member/student/hedar/Hedar_files/TestGO_files/TestGO_files/TestCodes/rosen.m

Inheritance

Declaration

class RosenbrockEvaluator

Description

Implementation of Rosenbrock Fucntion Source: http://www-optima.amp.i.kyoto-u.ac.jp/member/student/hedar/Hedar_files/TestGO_files/TestGO_files/TestCodes/rosen.m

Number of variables: variable (default is 2) Definition: f(x) = sum (i = 0 to n) [100 * ((x_i)^2 - (x_(i+1))^2)^2 + (x_i - 1)^2] Search domain: -2 <= xi <= 2, i = 1, 2,..., n. Number of local minima: several local minima. The global minima: x* = (1, ..., 1), f(x*) = 0. 

Public Member Functions

Type	Name	Description
override void	addDesignVariables ()	Adds design variables.
override void	addObjectives ()	Adds objectives.
override ModelEvaluation	EvaluateModel (object[] designVariables)	Evaluates function.
override bool	IsDesignAcceptable (object[] design, AreDesignsEqual areDesignsEqual)	Determines whether specified design is a local or global optima.
	RosenbrockEvaluator ()	Creates 2 dimensional Rosenbrock evaluator with bounds on the design variables.
	RosenbrockEvaluator (int numDimensions)	Creates N dimensional Rosenbrock evaluator with bounds on the design variables.
	RosenbrockEvaluator (int numDimensions, bool useDesignVariableBounds)	Creates N dimensional Rosenbrock evaluator with the ability to specify whether or not to bound design variables.
ModelEvaluation	EvaluateModel (object[] designVariables)	Evaluates the model at the given design point.
bool	IsDesignAcceptable (object[] design, AreDesignsEqual areDesignsEqual)	Determines whether design is an acceptable design.
void	SetStartPoint (object[] startValues)	Sets the starting design for the evaluator.

Public Member Functions inherited from ModelEvaluator

Type	Name	Description
void	addConstraint (string name, double lowerBound, double upperBound)	Method used to add a constraint
virtual void	addConstraints ()	Method called to specifiy the constraints for the evaluator (by default, not constraints are specified)
void	addDesignVariable (string name, object startValue)	Method used to add design variable
void	addDesignVariable (string name, object startValue, DataTable selectedAlphabet)	Method used to add design variable
void	addDesignVariable (string name, object startValue, double lowerBound, double upperBound)	Method used to add design variable
void	addDesignVariable (string name, object startValue, double lowerBound, double upperBound, DataTable selectedAlphabet)	Method used to add a design variable
abstract void	addDesignVariables ()	Method called to added design variables
void	addObjective (string name, double solveForValue)	Method used to add an objective (assumes solve for objective)
void	addObjective (string name, double solveForValue, double weight)	Method used to add an objective (assumes solve for objective)
void	addObjective (string name, Goal goal)	Method used to add an objective
void	addObjective (string name, Goal goal, double weight, double solveForValue)	Method used to add an objective
abstract void	addObjectives ()	Method called to specifiy the objectives for the evaluator
void	Dispose ()	Called to dispose the object.
abstract ModelEvaluation	EvaluateModel (object[] designVariables)	Evaluates the model at the given design point
virtual double	getIGD (List< double[]> bestDesignObjectives, string filePath)	Calculates the IGD value for multi-objective problem [The IGD formula is slightly modified to get the nearest possible value to actual IGD] [The actual formula accounts for equal no of optimal & obtained set of objectives, while, we calculate based on the number of objectives obtained]
abstract bool	IsDesignAcceptable (object[] design, AreDesignsEqual areDesignsEqual)	Determines whether design is an acceptable design.
void	SetStartPoint (object[] startValues)	Sets the starting design for the evaluator

Properties

Type	Name	Description
override object[,]	GlobalBestDesigns [get]	List of the globally best designs for the evaluator.
override bool	HasFailedRuns [get]	Evaluator does not have failed runs.
override bool	HasNonSmoothResponses [get]	Evaluator has smooth responses.
override string	Name [get]	Name of evaluator.
override bool	UsesConstraints [get]	Evaluator does not use constraints.
override bool	UsesDiscreteVariables [get]	Evaluator does not use multiple discrete variables.
override bool	UsesMinMax [get]	Evaluator does uses minimize/maximize objectives.
override bool	UsesMultipleObjectives [get]	Evaluator does not use multiple objectives.
override bool	UsesSolveFor [get]	Evaluator does not use the solve for objective.

Properties inherited from ModelEvaluator

Type	Name	Description
List< OptConstraint >	Constraints [get]	Method to get the list of constraints
int	DesignVariableCount [get]	Number of design variables specificed by the evaluator
List< DesignVariable >	DesignVariables [get]	Method to get the list of design variables
abstract object[,]	GlobalBestDesigns [get]	The global best design for the evaluator (a.k.a. "The right answers")
abstract bool	HasFailedRuns [get]	Does the evalutor have failed runs?
abstract bool	HasNonSmoothResponses [get]	Does the evaluator have non-smooth responses
abstract string	Name [get]	Name of the evaluator
int	NumberOfObjectives [get]	Number of objectives specified
List< Objective >	Objectives [get]	Method to get the list of objectives
abstract bool	UsesConstraints [get]	Does the evaluator use constraints?
abstract bool	UsesDiscreteVariables [get]	Does the evaluator use discrete variables?
abstract bool	UsesMinMax [get]	Does the evaluator use a Minimize/Maximize objective?
abstract bool	UsesMultipleObjectives [get]	Does the evaluator specify multiple objectives?
abstract bool	UsesSolveFor [get]	Does the evaluator use the solve for objective?

Properties inherited from IEvaluator

Type	Name	Description
List< OptConstraint >	Constraints [get]	List of constraints defined by the evaluator.
int	DesignVariableCount [get]	Number of design variables specificed by the evaluator.
List< DesignVariable >	DesignVariables [get]	List of design variables defined by the evaluator.
object[,]	GlobalBestDesigns [get]	The global best designs for the evaluator.
bool	HasFailedRuns [get]	Does the evalutor have failed runs?
bool	HasNonSmoothResponses [get]	Does the evaluator have non-smooth responses.
string	Name [get]	Name of the evaluator.
int	NumberOfObjectives [get]	Number of objectives specified.
List< Objective >	Objectives [get]	List of objectives defined by the evaluator.
bool	UsesConstraints [get]	Does the evaluator use constraints?
bool	UsesDiscreteVariables [get]	Does the evaluator use discrete variables?
bool	UsesMinMax [get]	Does the evaluator use a Minimize/Maximize objective?
bool	UsesMultipleObjectives [get]	Does the evaluator specify multiple objectives?
bool	UsesSolveFor [get]	Does the evaluator use the solve for objective?

Protected Member Functions inherited from ModelEvaluator

Type	Name	Description
virtual void	Dispose (bool disposing)	Standard disposal.

Constructor & Destructor Documentation

RosenbrockEvaluator()

RosenbrockEvaluator ( )

Creates 2 dimensional Rosenbrock evaluator with bounds on the design variables.

RosenbrockEvaluator()

RosenbrockEvaluator ( int numDimensions)

Creates N dimensional Rosenbrock evaluator with bounds on the design variables.

Parameters:

• numDimensions - Number of dimensions for the Rosenbrock evaluator.

RosenbrockEvaluator()

RosenbrockEvaluator ( int numDimensions, bool useDesignVariableBounds )

Creates N dimensional Rosenbrock evaluator with the ability to specify whether or not to bound design variables.

Parameters:

• numDimensions - Number of dimensions for the Rosenbrock evaluator.
• useDesignVariableBounds - True to bound design variables; false otherwise.

Member Function Documentation

addDesignVariables

override void `addDesignVariables` ( )

Adds design variables.

Implements ModelEvaluator.

addObjectives

override void `addObjectives` ( )

Adds objectives.

Implements ModelEvaluator.

EvaluateModel

override ModelEvaluation `EvaluateModel` ( object[] designVariables)

Evaluates function.

Parameters:

• designVariables - Design used to evaluate function.

Returns:

• Design results

Implements ModelEvaluator.

IsDesignAcceptable

override bool `IsDesignAcceptable` ( object[] design, AreDesignsEqual areDesignsEqual )

Determines whether specified design is a local or global optima.

Parameters:

• design - Design to be tested.
• areDesignsEqual - Function delegate to be be used to determine whether two designs are equal.

Returns:

• Returns true if the design is a local or global optima; false otherwise.

Implements ModelEvaluator.

Property Documentation

GlobalBestDesigns

override object [,] GlobalBestDesigns

List of the globally best designs for the evaluator.

Implements IEvaluator.

HasFailedRuns

override bool HasFailedRuns

Evaluator does not have failed runs.

Implements IEvaluator.

HasNonSmoothResponses

override bool HasNonSmoothResponses

Evaluator has smooth responses.

Implements IEvaluator.

Name

override string Name

Name of evaluator.

Implements IEvaluator.

UsesConstraints

override bool UsesConstraints

Evaluator does not use constraints.

Implements IEvaluator.

UsesDiscreteVariables

override bool UsesDiscreteVariables

Evaluator does not use multiple discrete variables.

Implements IEvaluator.

UsesMinMax

override bool UsesMinMax

Evaluator does uses minimize/maximize objectives.

Implements IEvaluator.

UsesMultipleObjectives

override bool UsesMultipleObjectives

Evaluator does not use multiple objectives.

Implements IEvaluator.

UsesSolveFor

override bool UsesSolveFor

Evaluator does not use the solve for objective.

Implements IEvaluator.