Evaluation of the Performance of Randomized FFD Control Grids for Design Optimization
Free-form deformation (FFD) is a commonly applied deformation method to solve real-world problems in design optimization like automotive shape optimization. The setup of FFD, which is the distribution of control points on a grid, has tremendous influence on the performance of the optimization process.
In a Bachelor thesis we want to numerically rate randomized control point distributions with given performance criteria to chose the best distribution for an shape optimization test scenario. Therefore, the task is to implement FFD and evaluate randomized FFD grids in an already implemented test scenario: 1D function approximation.
As a Master thesis the direct manipulation of FFD (DM-FFD) is to be implemented additionally as well as to be compared to the standard/indirect method in a further test scenario: 3D template fitting.
- Experience in C++
- Knowledge in computer graphics (Freeform curves/surfaces)
- Motivated to get insights into evolutionary optimization algorithms