Package org.moeaframework.util.weights
Class RandomGenerator
java.lang.Object
org.moeaframework.util.weights.RandomGenerator
- All Implemented Interfaces:
WeightGenerator
Generates weights randomly. This is the method proposed in [1] to replace the normal boundary intersection
method in the original MOEA/D. If
N weights are requested, this method first generates 50*N
random weights. From these weights, N are selected that are maximally distant from all other weights.
References:
- Zhang, Q., et al. "The Performance of a New Version of MOEA/D on CEC09 Unconstrained MOP Test Instances." IEEE Congress on Evolutionary Computation, 2009.
-
Constructor Summary
ConstructorsConstructorDescriptionRandomGenerator(int numberOfObjectives, int numberOfPoints) Constructs a new weight generator that generates randomly-sampled weights. -
Method Summary
-
Constructor Details
-
RandomGenerator
public RandomGenerator(int numberOfObjectives, int numberOfPoints) Constructs a new weight generator that generates randomly-sampled weights.- Parameters:
numberOfObjectives- the number of objectivesnumberOfPoints- the number of weights to generate
-
-
Method Details
-
size
public int size()Description copied from interface:WeightGeneratorReturns the number of weights that will be generated.- Specified by:
sizein interfaceWeightGenerator- Returns:
- the number of weights that will be generated
-
generate
Description copied from interface:WeightGeneratorReturns the generated weights.- Specified by:
generatein interfaceWeightGenerator- Returns:
- the generated weights
-