Within the scope of this thesis a flexible interface (Interface for Parametric Optimization, IPO) between the finite element solver Abaqus and the open source optimization library DAKOTA (Design Analysis Kit for Optimization and Terascale Applications) was developed. Finite element models created with Abaqus can be parameterized and optimized with respect to an arbitrary objective function and optional restrictions. Any mathematical combination of output variables available in Abaqus may serve as an objective function or restriction. DAKOTA provides a wide variety of different algorithms for optimization, parametric studies, uncertainty quantification and many other applications. Gradient based algorithms as well as gradient free methods, e.g., evolutionary strategies, can be chosen for solving the optimization problem. The IPO combines the advantages of both software packages. One can use the finite element solver Abaqus, which is capable of solving highly nonlinear (material as well as geometric nonlinearities) engineering problems and join it with the extensive optimization and parametric study capabilities of DAKOTA. The Abaqus Python application programming interface (API) serves as an easy--to--use basis for the coding, since all Abaqus pre-- and postprocessing commands are available in this API. An object oriented approach was chosen for the Interface for Parametric Optimization since is fits best into the Abaqus Python API and provides a convenient way for further extensions of the interface. The program was applied to the optimization of a simple truss construction and a more sophisticated bridge construction with their total weight as an objective function. The differences between several different optimization algorithms are then discussed in detail, highlighting their advantages and disadvantages.