Package org.moeaframework.problem.LSMOP

Interface CorrelationMatrix

public interface CorrelationMatrix
Correlation matrices used by the LSMOP test problem suite. The matrix describes the correlation between each decision variable group and the corresponding objective value.

A value of 0.0 indicates the decision variable group does not contribute to the objective value; whereas a value of 1.0 includes the decision variable group in the calculation. While it's possible for the matrix to contain values between [0, 1], the LSMOP suite does not use such intermediate values.

  • Field Summary

    Fields
    Modifier and Type
    Field
    Description
    static final CorrelationMatrix
    Full
    Creates a full correlation matrix where every decision variable group is included in the objective value calculation.
    static final CorrelationMatrix
    Overlapped
    Creates a correlation matrix where the i-th and (i+1)-th decision variable groups are associated with the i-th objective.
    static final CorrelationMatrix
    Separable
    Creates a correlation matrix with no correlation.

  • Method Summary

    Modifier and Type
    Method
    Description
    double[][]
    apply(int M)
    Returns an M x M correlation matrix containing values between [0, 1].

  • Field Details

    • Separable

      static final CorrelationMatrix Separable
      Creates a correlation matrix with no correlation. Each decision variable group is associated with a single objective.

    • Overlapped

      static final CorrelationMatrix Overlapped
      Creates a correlation matrix where the i-th and (i+1)-th decision variable groups are associated with the i-th objective.

    • Full

      static final CorrelationMatrix Full
      Creates a full correlation matrix where every decision variable group is included in the objective value calculation.

  • Method Details

    • apply

      double[][] apply(int M)
      Returns an M x M correlation matrix containing values between [0, 1].
      Parameters:
      M - the dimension
      Returns:
      the correlation matrix