Abstract Empirical Design of Geometric Algorithms.pdf

Empirical Design of Geometric Algorithms Karsten Weihe’ Uhik Brandesl Annegret Liebersl Matthias Miiller-Hannemann2 Dorothea Wagner1 Thomas Willhalml Abstract cles and the strategiesto tackle them were quite similar. It seems to us that this coincidence is not by chance,so a general, compar- The computer--aidedsolution to algorithmic problemsis becoming ative, retrospectiveanalysis of finished and on-going projectsmay moreand moreimportant in various application domains.This is in beworth the effort. particular true for computational geometry. For example,geomet- More specifically, we will restrict the discussionto the follow- ric problemsnaturally arise in image processing,computergraph- ing obstacles,which occurredin various ways: ics, and all kinds of computer-aideddesign,just to mention a few. Even more, the generaltendency towardsthe application of visual Formalization Obstacle: The algorithmic problemis too com- aidsin virtually all fields of science,technology,andbusinessraises plex to understandall relevantdetails andtheir interrelations. many new,unexpectedgeometricchallenges. Obvious reductionsto intelligible formulations seemto miss A sound mathematicaltreatmentof theseproblemsand a sys- crucial aspects. In the worst case,the outputs of an algo- tematic computational study on the resulting algorithms aredesir- rithm that