org.ejml.dense.row.linsol.qr.SolveNullSpaceQR_DDRM Maven / Gradle / Ivy
/*
* Copyright (c) 2009-2018, Peter Abeles. All Rights Reserved.
*
* This file is part of Efficient Java Matrix Library (EJML).
*
* Licensed under the Apache License, Version 2.0 (the "License");
* you may not use this file except in compliance with the License.
* You may obtain a copy of the License at
*
* http://www.apache.org/licenses/LICENSE-2.0
*
* Unless required by applicable law or agreed to in writing, software
* distributed under the License is distributed on an "AS IS" BASIS,
* WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
* See the License for the specific language governing permissions and
* limitations under the License.
*/
package org.ejml.dense.row.linsol.qr;
import org.ejml.data.DMatrixRMaj;
import org.ejml.dense.row.CommonOps_DDRM;
import org.ejml.dense.row.decomposition.qr.QRDecompositionHouseholderTran_DDRM;
import org.ejml.interfaces.SolveNullSpace;
/**
* Uses QR decomposition to find the null-space for a matrix of any shape if the number of
* singular values is known. WARNING: This only uses the first several rows in the input matrix. The rest are
* ignored.
*
* Solves for AT=QR and the last column in Q is the null space.
*
* @author Peter Abeles
*/
public class SolveNullSpaceQR_DDRM implements SolveNullSpace {
CustomizedQR decomposition = new CustomizedQR();
// Storage for Q matrix
DMatrixRMaj Q = new DMatrixRMaj(1,1);
/**
* Finds the null space of A
* @param A (Input) Matrix. Modified
* @param numSingularValues Number of singular values
* @param nullspace Storage for null-space
* @return true if successful or false if it failed
*/
@Override
public boolean process(DMatrixRMaj A , int numSingularValues, DMatrixRMaj nullspace ) {
decomposition.decompose(A);
if( A.numRows > A.numCols ) {
Q.reshape(A.numCols,Math.min(A.numRows,A.numCols));
decomposition.getQ(Q, true);
} else {
Q.reshape(A.numCols, A.numCols);
decomposition.getQ(Q, false);
}
nullspace.reshape(Q.numRows,numSingularValues);
CommonOps_DDRM.extract(Q,0,Q.numRows,Q.numCols-numSingularValues,Q.numCols,nullspace,0,0);
return true;
}
@Override
public boolean inputModified() {
return true;
}
/**
* Special/Hack version of QR decomposition to avoid copying memory and pointless transposes
*/
private static class CustomizedQR extends QRDecompositionHouseholderTran_DDRM {
@Override
public void setExpectedMaxSize( int numRows , int numCols ) {
this.numCols = numCols;
this.numRows = numRows;
minLength = Math.min(numCols,numRows);
int maxLength = Math.max(numCols,numRows);
// Don't delcare QR. It will use the input matrix for worspace
if( v == null ) {
v = new double[ maxLength ];
gammas = new double[ minLength ];
}
if( v.length < maxLength ) {
v = new double[ maxLength ];
}
if( gammas.length < minLength ) {
gammas = new double[ minLength ];
}
}
/**
* Modified decomposition which assumes the input is a transpose of the matrix
*/
@Override
public boolean decompose( DMatrixRMaj A_tran ) {
// There is a "subtle" hack in the line below. Instead of passing in (cols,rows) I'm passing in
// (cols,cols) that's because we don't care about updating everything past the cols
setExpectedMaxSize(A_tran.numCols, Math.min(A_tran.numRows,A_tran.numCols));
// use the input matrix for its workspace
this.QR = A_tran;
error = false;
for( int j = 0; j < minLength; j++ ) {
householder(j);
updateA(j);
}
return !error;
}
}
public DMatrixRMaj getQ() {
return Q;
}
}