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 ExtremalGeneralizedEigenvalueSolver
card
- 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.