ORD File Format

If you are licensed to use the ILOG CPLEX MIP optimizer, the ORD file format is available to indicate priority orders and branching directions for specific variables. Variables that are not given an explicit priority or that do not appear in an ORD file are assigned 0 (zero) priority. An ORD file begins with a NAME indicator record and ends with an ENDATA record.

Integer variables are specified, one per line, with an optional branching direction (UP or DN) beginning in column 2 and 3. Names begin in column 5 or beyond. The variable name and its priority must be separated by one or more blank spaces.

Here is an example of an ORD file:

NAME

x3 10

DN x5 5

UP x7

ENDATA

ORD files created using ILOG CPLEX Versions 2.1 or earlier used a fixed format in which the various data fields were limited to eight characters in length and restricted to specific columnar positions in each line. The extensions provided in the new ILOG CPLEX ORD file reader allow for more descriptive names and greater overall input flexibility. Most fixed-format ORD files conform to the new format. Any files that do not conform can be converted to the new format using the convert utility that comes with the standard ILOG CPLEX distribution. "Converting File Formats" in the ILOG CPLEX User's Manual explains how to use that utility.

Priorities for special ordered sets can be read from an SOS file. If there is no priority order for the SOSs, but one has been specified by an ORD file, a priority is derived for each SOS. The priority assigned to the set is the highest priority of all of its members. Priorities are derived for S1 and S2 sets.


Previous Page: MST File Format  Return to Top Next Page: REW File Format