download of program package and hungarian documentation

QOPT is a library to build and solve quadratically constrained quadratic programming (QCQP) problems. Its use is possible from high-level programming languages. We consider the QCQP problems in the following standard form:

where simmetric , furthermore, for all , is symmetric and . If  and  matrices are positive semidefinite, then the minimization is a convex optimization problem. We have to note that the global optimum is guaranteed in the case of convex programming problems, where all local optimal points are always optimal globally. In case of non-convex problems, we can not guarantee that our package can find a global or local optimal solution.

Because of their importance, two special classes of QCQP problems are mentioned here. In the linear programming (LP) problem all the quadratic terms of QCQP are missing, i.e. the problem can be written as 

In case the convex quadratic programming (QP) problem, the quadratic constraints are missing from QCQP and matrix Q in the objective function is positive semidefinite. The standard for of QP can be written as 

The use of the QPOPT library is described in its user's manual, which describes its functionality and illustrates the use of the library with an example driver program written in C.

János Fülöp, Room L511.
Research Group of Operations Research and Decision Systems
e-mail:
phone: (36-1) 279-6112, fax: (36-1) 209-5267

Csaba Mészáros, Room L514.
Research Group of Operations Research and Decision Systems
e-mail:
phone: (36-1) 279-6120, fax: (36-1) 209-5267