Downloading NEOS

The latest stable version of NEOS can be downloaded here:

Installing NEOS

There are three modes for running NEOS. Python 2.2 or higher is required for the client and solver modes, Python 2.3 or higher is required for running your own server.

Installing the Client Tools

The client tools provide a simple Python script, a Java GUI, and kestrel executables for submitting. optimization problems to NEOS. These are provided for your convenience, but they are not necessary for using NEOS as you can also submit problems to NEOS using email, the Web interface, or writing your own XML-RPC client. More information on submitting to NEOS can be found by selecting a solver from the solver list or by reading the NEOS Server FAQ.

After unzipping the neos zipfile or tarball, you can submit a problem to NEOS by creating an XML file describing your problem (see the XML-RPC link from any of the solver pages for more information on creating this file) and running NeosClient.py myxmlfile

To use the Java GUI, cd to the ClientTools directory and run: java -cp ./NeosClient.jar:./apache/xmlrpc-1.2-b1.jar NeosClient. The GUI uses the XML-RPC client library developed by apache.


Installing the Solver Tools

The Solver Tools are only necessary if you wish to add your own solver to the NEOS library.

After downloading and unzipping the NEOS server package, you will need to edit the SolverTools/config.py file to indicate where NEOS can creating working directories (NEOS will try to create them if they don't exist yet). The following variables need to be entered:
  NEOS_HOST="neos.mcs.anl.gov"
  NEOS_PORT=3332
  JOBSDIR=directory for jobs to run in
  LOGDIR=directory for logging jobs
  TESTDIR=directory for running tests
  DRIVER_FILE -- the location of the file that lists solvers available on your file system and where the drivers are located
(see the driver writing instructions section for more information).

AMPL and GAMS are the location of AMPL and GAMS executables.  These values are only necessary if any of your solvers will be using AMPL or GAMS AND you want to derive your solver drivers from the provided Python base classes.

@NEOS_WEB_DISCLAIMER@