Problem Names

Problems loaded in the Optimizer have a name. The name is either taken from the file name if the problem is read into the optimizer or it is specified as a string in a function call when a problem is loaded into the Optimizer using the library interface. Once loaded the name of the problem can be queried and modified. For example, the library provides the function XPRSsetprobname for changing the name of a problem.

When reading a problem from a matrix file the user can optionally specify a file extension. The search order used for matrix files in the case where the file extension is not specified is described in the reference for the function XPRSreadprob. In this case, the problem name becomes the file name, including the full path, but without the file extension.

Note that matrix files can be read directly from a gzip compressed file. Recognized names of matrix files stored with gzip compression have an extension that is one of the usual matrix file format extensions followed by the .gz extension. For example, hpw15.mps.gz.

The problem name is used as a default base name for the various file system interactions that the Optimizer may make when handling a problem. For example, when commanded to read a basis file for a problem and the basis file name is not supplied with the read basis command the Optimizer will try to open a file with the problem name appended with the .bss extension.

It is useful to note that the problem name can include file system path information. For example, c:/matrices/hpw15. Note the use of forward slashes in the Windows path string. It is recommended that Windows users use forward slashes as path delimiters in all file name specifications for the Optimizer since (i) this will work in all situations and (ii) it avoids any problems with the back slash being interpreted as the escape character.