Any additional information related to this implementation, that came to light after the release of this product, is described below.

1. Missing data files

   A few of the examples require data files which were inadvertantly omitted from the distribution materials. These are contained within the zip file mb24_data_files.zip (1,143 bytes). The contents should be unzipped and placed in the help\NAG\example directory within your NAG Toolbox installation, e.g. C:\Program Files\NAG\MB24\mbw6i24ddl\help\NAG\example.

   This table shows the data files contained within the zip file and the example programs which use them:

   Data file	Required by
   e04mx.opt	e04mx_example.m
   nag_opt_miqp_mps_read.opt	nag_opt_miqp_mps_read_example.m
   foo.data	x04bb_example.m,  nag_file_line_read_example.m

2. Fix for minor formatting problem in help

   For some classes of mathematical expressions, extraneous characters are displayed on the right hand side of the window following the proper mathematical display. The addition of the lines

   table.mtable{
   display: none;
   }
   

   to the end of the the file help\NAG\styles\libdoc.css within your NAG Toolbox for MATLAB distribution will hide these junk characters. Alternatively, you could replace that file by the one from our website: http://www.nag.co.uk/numeric/MB/manual64_24_1/html/styles/libdoc.css

3. MATLAB crashing when calling linear algebra routines

   If MATLAB is crashing when you call a linear algebra routine, this may be due to a problem with the DLL mkl_avx2.dll which is shipped with the NAG Toolbox. This file can be found in the C:\Program Files\NAG\MB24\mbw6i24ddl\mex.w64 directory (assuming the NAG Toolbox was installed in the default location). To work around the problem, rename the mkl_avx2.dll file. In the absence of this DLL, a different DLL is picked up instead, thus avoiding the problem. This will only affect machines which include the avx2 instruction set.

4. MATLAB R2014a Suitability

   This implementation has been tested with MATLAB version 8.3 (R2014a) and is considered suitable for use with that version.

5. MATLAB R2014b, R2015a, R2015b and R2016a Suitability

   This implementation has been tested with MATLAB version 8.4 (R2014b), version 8.5 (R2015a), version 8.6 (R2015b) and version 9.0 (R2016a) and is considered suitable for use with those versions. However, due to changes in the MATLAB graphics system at release R2014b, some of the example programs shipped with this implementation do not work correctly in these versions of MATLAB.