To comment on the following update, log in, then open the issue:
http://www.openoffice.org/issues/show_bug.cgi?id=50154
                  Issue #:|50154
                  Summary:|Numerical stability issue with the MINVERSE()
                          |function
                Component:|Spreadsheet
                  Version:|OOo 1.1.4
                 Platform:|All
                      URL:|
               OS/Version:|Windows XP
                   Status:|UNCONFIRMED
        Status whiteboard:|
                 Keywords:|
               Resolution:|
               Issue type:|DEFECT
                 Priority:|P3
             Subcomponent:|programming
              Assigned to:|spreadsheet
              Reported by:|masato12610





------- Additional comments from [EMAIL PROTECTED] Tue May 31 16:45:10 -0700 
2005 -------
I believe there is an issue with the numerical stability of the MINVERSE 
worksheet function.  From matrix algebra, if A is a square matrix (n rows and 
n columns) and is not singular, then a matrix inverse "Inv(A)" exists, such 
that for matrix multiplication A*Inv(A) = I, where I is the n-by-n identity 
matrix, i.e., the main diagonal of I contains ones and all other elements of I 
contain zero.

Attached are Excel and Calc workbooks that illustrate the issue.  The Excel 
workbook appears to work correctly but the Calc workbook shows the issue.

In the Excel workbook there is a Worksheet "Example1" that has a 10X10 matrix 
(called MatrixA) at the top. The next matrix down is the inverse of the first 
matrix (InvMatrixA) calculated with the MINVERSE() function.  Below the 
inverse is the matrix product calculated with the MMULT() function, i.e., MMULT
(MatrixA,InvMatrixA), which should result in the identify matrix.  

Example1 shows correct numerical results.  The main diagonal has "1".  The 
cells of the main diagonal are highlighted "yellow".  All other cells are 
essentailly "zero", i.e. on the order of 1E-13 or less.

Example2 worksheet is structured similarily with an 18X18 matrix.  The bottom 
matrix shows ones on the main diagonal with the other cells close to zero as 
well, i.e., close to being an identity matrix.

In the Calc version of you'll see that Example1 shows correct numerical 
results.

However, in Example2, the bottom matrix is not "close" to being an identity 
matrix.

This problem shows up in both OOo 1.1.4 and 2.0Beta.

---------------------------------------------------------------------
Please do not reply to this automatically generated notification from
Issue Tracker. Please log onto the website and enter your comments.
http://qa.openoffice.org/issue_handling/project_issues.html#notification

---------------------------------------------------------------------
To unsubscribe, e-mail: [EMAIL PROTECTED]
For additional commands, e-mail: [EMAIL PROTECTED]


---------------------------------------------------------------------
To unsubscribe, e-mail: [EMAIL PROTECTED]
For additional commands, e-mail: [EMAIL PROTECTED]

Reply via email to