FortMP-MEX add-on to Matlab is an optimisation solver system capable of solving large-scale industrial applications ranging from supply chain to portfolio planning and asset liability management in finance. The system is designed to solve a wide range of well known classes of mathematical programming problems including large-scale linear, variable separable, mixed integer, quadratic and quadratic mixed integer problems.
Linear programming problems are processed by the sparse simplex (SSX) or interior point method (IPM). A powerful basis recovery (cross over) algorithm combines the speed of the IPM solution with the warm restart property of the SSX. Mixed integer programs are solved by applying a branch and bound tree search method. Activating cutting plane methods and integer pre-processing techniques will make FortMP-MEX highly competitive and effective for MIP problems.
Matlab environment provides a platform for direct interaction with the solving capabilities of FortMP. A user can build a model in a regular M-file and invoke FortMP functions to obtain a solution. The software also supports the MPS file format. All solutions are transferred into the Matlab workspace and solver controls can be set either by simple command line arguments or by special solver control file called a spec file. Large-scale models which could not be otherwise solved, can be processed by the Matlab optimisation toolbox.
FortMP provides an industrial strength optimiser, which under the Matlab environment provide a versatile way of solving problems by offering various alternatives of specifying a problem. In addition it provides a powerful and desirable alternative for both large-scale sparse models and for discrete (integer) decision models.
It will provide a very transparent means for Matlab users to interact with an industrial strength solver. In this way it will exploit a niche means for researchers and investors or all Matlab users in general who need to develop and solve complex problems in the comfort of the Matlab environment. The problems that can be solved include Integer Programming (IP), Mixed Integer Programming (MIP), Quadratic Programming (QP) and Quadratic Mixed Integer Programming (QMIP). These are not all currently supported in the existing Matlab solvers. MATLAB Key features include:
- Solve Large scale Linear Programming (LP) Problems using SSX and Barrier/IPM
- Solve Large Scale Mixed Integer Programming (MIP) Problems
- Solve Quadratic Programming (QP) Problems
- Solve Quadratic Mixed Integer Programming (QMIP) Problems
- Customisable Solver Control
- Used in Portfolio Optimisation and Asset Liability Management
- All Windows versions
Since we are constantly working on new releases of FortMP-MEX, please contact OptiRisk Systems for updated information.
Example of Use
To see the of FortMP-MEX example of use please click here.