Package com.polytechnik.kgo
Class KGOIterationalSimpleOptimizationU
java.lang.Object
com.polytechnik.kgo.KGOIterationalSimpleOptimizationU
A Knowledge Generalized Operator iteration optimization,
directly in u_orig basis.
The algorithm uses an adjustment from simplified to full version of constraints:
\( \sum\limits_{k=0}^{nX-1}u_{ik}u_{jk}=\delta_{ij} \), for \(i,j=0\dots nC-1\).
The Lagrange multipliers are calculated with
LagrangeMultipliersPartialSubspace.calculateRegularLambda(int, int, double[], double[])
. With this
\( \lambda_{ij} \) the algorithm does not converge well, but a good
enough solution can be found.
A try without SVD.- See Also:
-
Field Summary
FieldsModifier and TypeFieldDescriptionfinal UAdjustment
final KGOEVSelection
private static final boolean
private static final int
-
Constructor Summary
Constructors -
Method Summary
-
Field Details
-
N_iterations
private static final int N_iterations- See Also:
-
FORCE_DIAG_DUMP
private static final boolean FORCE_DIAG_DUMP- See Also:
-
evSelected
-
aep
-
-
Constructor Details
-
KGOIterationalSimpleOptimizationU
public KGOIterationalSimpleOptimizationU(int nC, int nX, double[] SK, double eps)
-