Case-based tuning of a metaheuristic algorithm exploiting sensitivity analysis and design of experiments for reverse engineering applications

Due to its capacity to evolve in a large solution space, the Simulated Annealing (SA) algorithm has shown very promising results for the Reverse Engineering of editable CAD geometries including parametric 2D sketches, 3D CAD parts and assemblies. However, parameter setting is a key factor for its performance, but it is also awkward work. This paper addresses the way a SA-based Reverse Engineering technique can be enhanced by identifying its optimal default setting parameters for the fitting of CAD geometries to point clouds of digitized parts. The method integrates a sensitivity analysis to characterize the impact of the variations in the parameters of a CAD model on the evolution of the deviation between the CAD model itself and the point cloud to be fitted. The principles underpinning the adopted fitting algorithm are briefly recalled. A framework that uses design of experiments (DOEs) is introduced to identify and save in a database the best setting parameter values for given CAD models. This database is then exploited when considering the fitting of a new CAD model. Using similarity assessment, it is then possible to reuse the best setting parameter values of the most similar CAD model found in the database. The applied sensitivity analysis is described together with the comparison of the resulting sensitivity evolution curves with the changes in the CAD model parameters imposed by the SA algorithm. Possible improvements suggested by the analysis are implemented to enhance the efficiency of SA-based fitting. The overall approach is illustrated on the fitting of single mechanical parts but it can be directly extended to the fitting of parts' assemblies. It is particularly interesting in the context of the Industry 4.0 to update and maintain the coherence of the digital twins with respect to the evolution of the associated physical products and systems.