public class ExtremalGeneralizedEigenvalueBySDP extends Object implements ExtremalGeneralizedEigenvalueSolver
| Constructor and Description |
|---|
ExtremalGeneralizedEigenvalueBySDP(SymmetricMatrix A,
SymmetricMatrix B)
Constructs the problem described in Section 3.2, d'Aspremont (2008),
changed to a minimization problem.
|
ExtremalGeneralizedEigenvalueBySDP(SymmetricMatrix A,
SymmetricMatrix B,
boolean isMinimizationProblem)
Constructs the problem described in Section 3.2, d'Aspremont (2008).
|
ExtremalGeneralizedEigenvalueBySDP(SymmetricMatrix A,
SymmetricMatrix B,
int maxIterations,
double weightRank,
double weightCardinality,
double tol,
boolean isMinimizationProblem)
Constructs the problem described in Section 3.2, d'Aspremont (2008).
|
| Modifier and Type | Method and Description |
|---|---|
Vector |
computeOptimalPositions(int card)
Computes the solution to the problem described in Section 3.2 in reference.
|
double |
getLambda(Vector x)
Computes the value of the objective function in eq.
|
public ExtremalGeneralizedEigenvalueBySDP(SymmetricMatrix A, SymmetricMatrix B)
A - the matrix A in eq. 10, reference [1]B - the matrix B in eq. 10, reference [1]public ExtremalGeneralizedEigenvalueBySDP(SymmetricMatrix A, SymmetricMatrix B, boolean isMinimizationProblem)
A - the matrix A in eq. 10, reference [1]B - the matrix B in eq. 10, reference [1]isMinimizationProblem - if true, eq. 10 in reference [1] is
changed to a
minimization problempublic ExtremalGeneralizedEigenvalueBySDP(SymmetricMatrix A, SymmetricMatrix B, int maxIterations, double weightRank, double weightCardinality, double tol, boolean isMinimizationProblem)
A - the matrix A in eq. 10, reference [1]B - the matrix B in eq. 10, reference [1]maxIterations - the maximum number of iterationsweightRank - the penalty weight for rank constraintsweightCardinality - the penalty weight for cardinality constraintstol - the tolerance for rank and cardinality constraintsisMinimizationProblem - if true, eq. 10 in reference [1] is
changed to a
minimization problempublic Vector computeOptimalPositions(int card)
computeOptimalPositions in interface ExtremalGeneralizedEigenvalueSolvercard - the cardinality constraint, the number of non-zero elements in x, the solutionpublic double getLambda(Vector x)
x - variable x in eq. 10, reference[1]Copyright © 2010-2020 NM FinTech Ltd.. All Rights Reserved.