C# Class SwarmOps.Optimizers.MetaFitness

Compute the standard Meta-Fitness measure, that is, perform a number of optimization runs on different problems and sum their results.

Preemptive Fitness Evaluation is used in that optimizations will be attempted aborted once the meta-fitness becomes worse than the Preemptive Fitness Limit (aka. fitnessLimit). In addition, the array of problems are being sorted at the end of each meta-fitness computation, so as to allow the worst performing problems to be optimized first in the next meta-fitness computation, so as to decrease the computation time further still.

Inheritance: Problem

Exibir arquivo Open project: DanWBR/dwsim3

Protected Properties

Property	Type	Description
ProblemIndex	ProblemIndex

Public Methods

Method	Description
BeginOptimizationRun ( ) : void	At beginning of new meta-optimization run print a newline.
EnforceConstraints ( double &parameters ) : bool	Enforce constraints and evaluate feasiblity.
Feasible ( double parameters ) : bool	Evaluate feasibility (constraint satisfaction).
Fitness ( double parameters, double fitnessLimit ) : double	Compute the meta-fitness measure by passing the given parameters to the Optimizer, and perform optimization runs on the array of problems until the fitness compute exceeds the fitnessLimit.
MetaFitness ( Optimizer optimizer, Problem problems, int numRuns, int maxIterations ) : System.Diagnostics	Construct the object, un-weighted problems.
MetaFitness ( Optimizer optimizer, WeightedProblem weightedProblems, int numRuns, int maxIterations ) : System.Diagnostics	Construct the object, weighted problems.

Method Details

BeginOptimizationRun() public method

At beginning of new meta-optimization run print a newline.

public BeginOptimizationRun ( ) : void
return	void

EnforceConstraints() public method

Enforce constraints and evaluate feasiblity.

public EnforceConstraints ( double &parameters ) : bool
parameters	double	Parameters to use for the Optimizer.
return	bool

Feasible() public method

Evaluate feasibility (constraint satisfaction).

public Feasible ( double parameters ) : bool
parameters	double	Parameters to use for the Optimizer.
return	bool

Fitness() public method

Compute the meta-fitness measure by passing the given parameters to the Optimizer, and perform optimization runs on the array of problems until the fitness compute exceeds the fitnessLimit.

public Fitness ( double parameters, double fitnessLimit ) : double
parameters	double	Parameters to use for the Optimizer.
fitnessLimit	double	Preemptive Fitness Limit
return	double

MetaFitness() public method

Construct the object, un-weighted problems.

public MetaFitness ( Optimizer optimizer, Problem problems, int numRuns, int maxIterations ) : System.Diagnostics
optimizer	Optimizer	Optimizer to be used.
problems	Problem	Array of problems to be optimized.
numRuns	int	Number of optimization runs per problem.
maxIterations	int	Max number of optimization iterations.
return	System.Diagnostics

MetaFitness() public method

Construct the object, weighted problems.

public MetaFitness ( Optimizer optimizer, WeightedProblem weightedProblems, int numRuns, int maxIterations ) : System.Diagnostics
optimizer	Optimizer	Optimize to be used.
weightedProblems	WeightedProblem	Array of weighted problems to be optimized.
numRuns	int	Number of optimization runs per problem.
maxIterations	int	Max number of optimization iterations.
return	System.Diagnostics

Property Details

ProblemIndex protected_oe property

Sorted index of optimization problems.

protected ProblemIndex ProblemIndex
return	ProblemIndex