ddsip

The NEOS Server offers DDSIP for the solution of mixed-integer stochastic linear programming problems with input in MPS format, or in LP format.

DDSIP implements a number of scenario decomposition algorithms for stochastic linear programs with mixed-integer recourse. Main idea is the Lagrangian relaxation of the nonanticipativity constraints and a branch and bound algorithm to reestablish nonanticipativity. For dual optimization it uses the ConicBundle algorithm of Christoph Helmberg. Mixed-integer subproblems are solved with the CPLEX callable library.

DDSIP supports mean-risk models involving the risk measures expected shortfall below target, excess probabilities, absolute semideviation, worst-case-costs, tail value-at-risk, and standard deviation. Any found bugs should be reported to the authors.

DDSIP was developed by Claus C. Carøe, Ralf Gollmer. Andreas Märkert, and Rüdiger Schultz.

This solver was implemented by Hans Mittelmann and executes at under

Using the NEOS Server for DDSIP

The user must submit a mixed-integer stochastic linear programming problem with a number of files, see this manual. When submitting via e-mail or XML-RPC delete empty tokens!

Enter the complete path to the specificationfile
Specification file (some CPLEX and branch and bound parameters):

Enter the complete path to the model file (LP)
Single-scenario model file (LP):

Enter the complete path to the model file (MPS)
Single-scenario model file (MPS):

Enter the complete path to the subproblem priority file (optional)
Priority order file for subproblems (optional):

Enter the complete path to the rhs scenario file
Right-hand side scenario file:

Enter the complete path to the cost scenario file (optional)
Cost scenario file (optional):

Enter the complete path to the matrix scenario file (optional)
Matrix scenario file (optional):

Enter the complete path to the master priority file (optional)
Priority order file for master (optional):

Enter the complete path to the start information file (optional)
Start information file (optional):

Comments:

e-mail address:

Please do not click the 'Submit to NEOS' button more than once.

_{DOE disclaimer

DOE Web privacy policy}