org.moeaframework.core.indicator
Class RIndicator

java.lang.Object
  org.moeaframework.core.indicator.NormalizedIndicator
      org.moeaframework.core.indicator.RIndicator

All Implemented Interfaces:

Indicator

Direct Known Subclasses:

R1Indicator, R2Indicator, R3Indicator

public abstract class RIndicator
extends NormalizedIndicator

Abstract class for implementing R indicator functions.

References:

1. Hansen, M. P. and A. Jaszkiewicz (1998). Evaluating the Quality of Approximations to the Non-dominated Set. IMM Technical Report IMM-REP-1998-7.

Nested Class Summary

static class	RIndicator.ChebychevUtility Chebychev (also referred to as Tchebycheff) utility function.
static class	RIndicator.LinearWeightedSumUtility Utility computed as the sum of the weighted objective values.
static interface	RIndicator.UtilityFunction Interface for defining utility functions.

Field Summary

protected RIndicator.UtilityFunction	utilityFunction The utility function used by the R2 calculation.
protected double[][]	weights The weights, typically uniformly distributed.

Fields inherited from class org.moeaframework.core.indicator.NormalizedIndicator

problem

Constructor Summary

RIndicator(Problem problem, int subdivisions, NondominatedPopulation referenceSet)
Constructs a new R indicator using the Chebychev utility function.

RIndicator(Problem problem, int subdivisions, NondominatedPopulation referenceSet, RIndicator.UtilityFunction utilityFunction)
Constructs a new R indicator using the specified utility function.

Method Summary

double	expectedUtility(NondominatedPopulation population) Computes the expected utility for the given population.
static int	getDefaultSubdivisions(Problem problem) Returns the default number of subdivisions for a given problem.

Methods inherited from class org.moeaframework.core.indicator.NormalizedIndicator

getNormalizedReferenceSet, normalize

Methods inherited from class java.lang.Object

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

Methods inherited from interface org.moeaframework.core.Indicator

evaluate

Field Detail

utilityFunction

protected RIndicator.UtilityFunction utilityFunction

The utility function used by the R2 calculation.

weights

protected double[][] weights

The weights, typically uniformly distributed.

Constructor Detail

RIndicator

public RIndicator(Problem problem,
                  int subdivisions,
                  NondominatedPopulation referenceSet)

Constructs a new R indicator using the Chebychev utility function.

Parameters:

problem - the problem

subdivisions - the number of subdivisions along each objective

referenceSet - the reference set

RIndicator

public RIndicator(Problem problem,
                  int subdivisions,
                  NondominatedPopulation referenceSet,
                  RIndicator.UtilityFunction utilityFunction)

Constructs a new R indicator using the specified utility function.

Parameters:

problem - the problem

subdivisions - the number of subdivisions along each objective

referenceSet - the reference set

utilityFunction - the utility function

Method Detail

expectedUtility

public double expectedUtility(NondominatedPopulation population)

Computes the expected utility for the given population.

Parameters:

population - the population

Returns:

the expected utility

getDefaultSubdivisions

public static int getDefaultSubdivisions(Problem problem)

Returns the default number of subdivisions for a given problem. The defaults, for an M objective problem, are:

• if M=2, then 500
• if M=3, then 30
• if M=4, then 12
• if M=5, then 8
• else 3

Parameters:

problem - the problem

Returns:

the default number of subdivisions for a given problem