[ User's Guide | Download | Bottom of page ]
IBM Optimization Library Stochastic Extensions
Introduction
Stochastic programming supports decision making under uncertainty. It is
a methodology for bringing uncertain future scenarios into the traditional
decision making framework of linear programming. Just as linear programming
models the optimal allocation of constrained resources to meet known demands,
stochastic programming models the allocation of today's resources to meet
tomorrow's unknown demands in such a way that the user can explore the
trade offs with respect to expected risks and rewards and make informed
decisions.
Users seeking the greater modeling power and flexibility of stochastic
programming come up against high hurdles: there is more input data to be
managed, the optimization problems are very large, and the solution arrays
are larger and more difficult to analyze.
IBM Stochastic Extensions offers advanced new technology to help the
user meet the demands of modeling very large, multi-stage stochastic programs.
While the IBM Stochastic Solutions product
has an easy-to-use solver that operates from the command line, as well as a
starter set of callable modules, Stochastic
Extensions provides advanced programming capability in conjunction
with the IBM Optimization Library. It supports the
Stochastic Mathematical Programming
System (SMPS) input format for multistage stochastic programs as well as
accepting model data from the application through a set of friendly
yet robust stochastic data
structures. The flexible nested
decomposition solver is robust and fast, and allows the developer total
control over partitioning large master problems into smaller sub-problems.
The IBM Stochastic Extensions requires the IBM Optimization Library
to provide the underlying solver capabilities. To use the flexible nested
decomposition solver, TCP/IP must also be installed and running on the
user's system.
The Stochastic Extensions Subroutine Modules
Highly advanced modeling development is supported by the IBM
Stochastic Extension Subroutine Modules. These are designed to facilitate
the integration of IBM Stochastic Extensions with IBM Optimization Library
applications. These subroutine modules are accessed by constructing specialized
driver programs. These functions
support advanced model generation, control of solver operation through
callbacks, and analysis of solution data. Input modules support in-core
modeling of stochastic programs, user callbacks monitor solver progress,
and data access modules retrieve the solution for use in application processing.
Guide, Reference and Examples
The Guide and Reference is divided into three parts. The User's
Guide contains basic information to help users formulate, solve, and
analyze stochastic programming problems. The Reference
Section contains detailed information about calling sequences, data
structures, and messages. The Examples
contains example problems and several driver programs.
Readers seeking a fuller introduction to the ideas and techniques of
stochastic programming should consult the recently published texts by Birge
and Louveaux, and Kall and
Wallace .
If you would like to have softcopy of the html user guide on your PC, click here to download the zipped postscript HTML (179 KB)
If you prefer hardcopy documentation, click here to download a postscript (1100KB) or zipped postscript (294KB) version.
| Notes: |
- Viewing these postcript files is problematical. An inability to view them satisfactorily does not necessarily mean that they will not print correctly.
- There are no hypertext links in the postcript files, and so the cross references are undetectable.
|
[
Top of Page |
Previous Page |
Stochastic Extensions User's Guide ]
