A Roadmap to Optimization Software Tools for High Performance Computing Environments

This roadmap is a service to the research communities working in both optimization and parallel computation. It is a survey of optimization algorithms (discrete optimization) implemented on either massively parallel (or vector-based) machines or clusters of high speed workstations. For each implementation, this roadmap will include the following items:

Type of the combinatorial algorithm(s) involved in the parallelization.
The problem(s) used in the implementation.
The type of hardware as well as number of processors used in the application.
References to other web pages to obtain more info about the implementation, methods involved, etc.
If available, a short description of the software and whether it is available upon request for research purposes or is it only commercial.
If available, how the results compare with results of non-parallel implementations.
The motivation or reasons for parallelizing the method (i.e.is it inherently parallel? etc.)
Contact addresses and emails of the people in charge of the implementation.

The tree structure of this roadmap follows the format of the NEOS Guide Optimization Tree . The implementations list is divided into subsections and each subsection is for one particular type of algorithm.

Integer Programming methods:
- Branch-and-Bound based methods software list.
- More Branch-and-Bound based methods software list.
Non-Integer Programming methods:
- Neural nets based methods software list.
- Simulated annealing based methods software list.
- Genetic algorithms based methods software list.
- Tabu search based methods software list.
- Dynamic programming based methods software list.

Other large scale optimization software and relevant pointers can be found at: