TOMS425
Generation of Random Correlated Normal Variables

TOMS425 is a FORTRAN77 library which implements ACM TOMS algorithm 425, for generating random correlated normal variables.

The code posted here has been updated to reflect a correction that saves the variable G02 in RNOR between calls, and other updates related to a remark by Page in June, 1974, supplied by Tim Hopkins, 24 January 2011.

The text of many ACM TOMS algorithms is available online through ACM: http://www.acm.org/pubs/calgo or NETLIB: http://www.netlib.org/toms/index.html.

Usage:

call rnvr ( x, a, nv, ni, ient, iarg )

• x contains the output normal variables;
• a is the covariance matrix;
• nv is the order of the covariance matrix;
• ni is the order of the array;
• ient is -1 to indicate the initial call;
• iarg is a seed for the random number generator.

Languages:

TOMS425 is available in a FORTRAN77 version.

Reference:

1. Rex Hurst, Robert Knop,
   Algorithm 425: Generation of Random Correlated Normal Variables,
   Communications of the ACM,
   Volume 15, Number 5, May 1972, pages 355-357.
2. R L Page,
   Remark on Algorithm 425: Generation of Random Correlated Normal Variables,
   Communications of the ACM,
   Volume 17, Number 6, June 1974, page 325.

Source Code:

• src.f, the source code.
• Makefile, a makefile.
• toms425.csh, commands to compile the source code.

Examples and Tests:

• driver.f, a sample calling program.
• toms425_prb.csh, commands to compile and run the sample program.
• Res, the output file.

List of Routines:

• RNVR generates a random correlated normal variable on each call.
• RNOR generates a random normal number on each call.

You can go up one level to the FORTRAN77 source codes.