The application of Mathematica for large-scale nonlinear optimization. Known software company KNITRO Ziena Optimization for large-scale nonlinear optimization is now released as an application program Mathmatica, which allows the use of algorithms directly in projects that run on Mathematica. KNITRO provides a wide range of custom options, and offers interfaces C, C, Fortran, Java, AMPL, AIMMS, GAMS, MPL, Mathematica, MATLAB Excel Microsoft, and LabVIEW. The company Softline - a major supplier of software for scientific, engineering and education. The company is the exclusive distributor in Russia of such well-known manufacturers of scientific software such as MathWorks (MATLAB, Simulink), Waterloo Maple (Maple), etc.
March 14, 2006 Company Wolfram Research has announced the application KNITRO program Mathematica, designed for large-scale nonlinear optimization. By combining advanced algorithms with the reliability of Mathematica, a new application KNITRO for Mathematica is an excellent tool for solving a wide range of tasks, from the development of business processes and production management to the optimization of integrated circuits.
Nonlinear Optimization - the basic method of designing in many areas, particularly applicable in industry, aerospace, industrial automation, business and military planning and logistics. Appendix KNITRO for Mathematica finds a local minimum for problems with continuous variables and smooth functions. It solves the problem with nonlinear constraints, linear programming, quadratic programming problem without constraints, and solved by the least squares method and a system of nonlinear equations.
Key Features: Newton methods, including BFGS, SR1 and L-BFGS, which quickly converge to the exact local solutions.
There is an option, which requires at each iteration step, the constraints (given by inequalities) and the domain of the variables.
Rapid transition from an interior point algorithm to the algorithm of the active set, and vice versa to determine the vertices.
Linear algebra operations for the iterative (conjugate gradient) and direct (sparse decomposition) methods.
Automatic selection of a starting point.
"We are encouraged by this new partnership with Wolfram Research," - said Richard Valts, president Ziena Optimization, developed the KNITRO for Mathematica, - It will provide access to users of Mathematica to modern optimization algorithms KNITRO from among the program Mathematica. Mathematica users will be able to solve large complex nonlinear optimization problems that are not available to other programs. "
Appendix KNITRO for Mathematica developed by Ziena Optimization, it requires for its operation Mathematica version 5.1 or higher for Windows and Linux, there are also versions for Sun Solaris and some Linux x86 (64-bit) platforms.