Author: luc
Date: Thu Dec 4 15:56:17 2008
New Revision: 723500
URL: http://svn.apache.org/viewvc?rev=723500&view=rev
Log:
moved the various solve function out of decomposition algorithms
and into a dedicated DecompositionSolver class
Modified:
commons/proper/math/trunk/src/site/xdoc/userguide/linear.xml
Modified: commons/proper/math/trunk/src/site/xdoc/userguide/linear.xml
URL:
http://svn.apache.org/viewvc/commons/proper/math/trunk/src/site/xdoc/userguide/linear.xml?rev=723500&r1=723499&r2=723500&view=diff
==============================================================================
--- commons/proper/math/trunk/src/site/xdoc/userguide/linear.xml (original)
+++ commons/proper/math/trunk/src/site/xdoc/userguide/linear.xml Thu Dec 4
15:56:17 2008
@@ -42,9 +42,7 @@
<ul>
<li>Matrix addition, subtraction, multiplication</li>
<li>Scalar addition and multiplication</li>
- <li>Inverse and transpose</li>
- <li>Determinants and singularity testing</li>
- <li>LU decomposition</li>
+ <li>transpose</li>
<li>Norm and Trace</li>
<li>Operation on a vector</li>
</ul>
@@ -67,8 +65,9 @@
System.out.println(p.getRowDimension()); // 2
System.out.println(p.getColumnDimension()); // 2
-// Invert p
-RealMatrix pInverse = new LUDecompositionImpl(p).inverse();
+// Invert p, using LU decomposition
+DecompositionSolver solver = new DecompositionSolver(p);
+RealMatrix pInverse = solver.inverse(solver.luDecompose());
</source>
</p>
</subsection>
@@ -94,7 +93,7 @@
</subsection>
<subsection name="3.4 Solving linear systems" href="solve">
<p>
- The <code>solve()</code> methods of the
<code>DecompositionSolver</code> interface
+ The <code>solve()</code> methods of the
<code>DecompositionSolver</code> class
support solving linear systems of equations of the form AX=B, either
in linear sense
or in least square sense. In each case, the <code>RealMatrix</code>
represents the
coefficient matrix of the system. For example, to solve the linear
system
@@ -103,19 +102,20 @@
-x + 7y + 6x = -2
4x - 3y - 5z = 1
</pre>
- Start by creating a RealMatrix to store the coefficients matrix A
+ Start by creating a RealMatrix to store the coefficients matrix A
and provide it
+ to a solver
<source>
RealMatrix coefficients =
new RealMatrixImpl(new double[][] { { 2, 3, -2 }, { -1, 7, 6 }, { 4, -3,
-5 } },
false);
+DecompositionSolver solver = new DecompositionSolver(coefficients);
</source>
Next create a <code>RealVector</code> array to represent the constant
- vector B and use <code>solve(RealVector)</code> from one of the
implementations
- of the <code>DecompositionSolver</code> interface, for example
- <code>LUDecompositionImpl</code> to solve the system
+ vector B, select one of the decomposition algorithms (LU, QR ...)
and use
+ <code>solve(vector, decomposition)</code> to solve the system
<source>
RealVector constants = new RealVectorImpl(new double[] { 1, -2, 1 }, false);
-RealVector solution = new LUDecompositionImpl(coefficients).solve(constants);
+RealVector solution = solver.solve(constants, solver.luDecomposition());
</source>
The <code>solution</code> vector will contain values for x
(<code>solution.getEntry(0)</code>), y
(<code>solution.getEntry(1)</code>),
@@ -146,7 +146,7 @@
It is possible to solve multiple systems with the same coefficient
matrix
in one method call. To do this, create a matrix whose column
vectors correspond
to the constant vectors for the systems to be solved and use
- <code>solve(RealMatrix),</code> which returns a matrix with column
+ <code>solve(matrix, decomposition),</code> which returns a matrix
with column
vectors representing the solutions.
</p>
</subsection>