PPRN

  1. Introduction.
  2. Additional documentation.
  3. Usage conditions.
  4. Getting PPRN.
  5. Using PPRN.

  1. Introduction.

    The PPRN package was developed in the Statistics and Operations Research Dept. at Universitat Politecnica de Catalunya, Barcelona (Spain), by Jordi Castro and Narcís Nabona. The package is appropriate for solving single/multicommodity network flow problems with linear/nonlinear objective function and with/without linear side constraints. Thus it can be viewed as a general package for solving a high variety of network flow problems (though it was originally designed for solving nonlinear multicommodity problems with side constraints).

    The code has been tested comparing its performance against other codes. When solving pure linear network flow problems, it has proved to be faster than alternative codes like NETFLO (J. Kennington) or RELAX-IV (D. Bertsekas) in all cases tried of a battery of dimacs problems. When solving linear multicommodity problems its performance was in general better, though not in all the cases tested, than that of MCNF85 (J. Kennington). For nonlinear problems it has only been compared with general purpose packages like MINOS (Murtagh and Saunders) and PPRN always outperformed it.

  2. Additional documentation.

  3. Usage conditions.

    You can only use the package for academic purposes. For other types of usage, please contact with the authors sending a message to jordi.castroupc.edu.

  4. Getting PPRN.

    PPRN is available in object form (no source code is provided). We have precompiled libraries for the following architectures (note that all these files are gzipped libraries, so yo have to gunzip them):

    Once you got and gunzip the appropriate file, rename it as libpprn.a. Remember to run "ranlib libpprn.a" after copying the library to the directory in your machine (when necessary).

    If you have a different architecture from those currently available, please contact us at jordi.castroupc.edu.

  5. Using PPRN.

    Follow these steps to test PPRN:

    If you have some comments or suggestions, or you find any trouble or abnormal situations when using the package, we would appreciate if you would send a message to jordi.castroupc.edu .