Title: Geometrical system decomposition using a multigroup approach
1Geometrical system decompositionusing a
multi-group approach
- P. Mathis, P. Schreck
- LSIIT Université Louis Pasteur
- UMR CNRS 7005
- Strasbourg - France
22D dimensioned sketch a constraint system
3Decomposition
- Goal cut the constraint system in smaller ones
- Because
- With resolution approach time is shorter
- With construction approach more problems are
solved - Rely on invariance under displacements
- Idea considering others groups
4First example
decomposable
graph not decomposable
but not well constrained modulo the displacement
group
(well constrained modulo the similarity group)
5Second example
6Constraint typology
translation
displacement
similarity
transformation
7Similarity invariant part
8Sub-figure 1similarities invariance
9Sub-figure 1similarities invariance
Solution 1 orbit of sub-figure 1 under
similarities
10Displacement invariant part
11Sub-figure 2Displacements invariance
12Assembling subf1 and subf2
Similarity computation
13New constraints extracted border
14Remaining problemTranslation invariance
15Sub-figure 3Translation invariance
16Assembling
Displacement computation
17Key points
- Hierarchy of groups
- G-reference entities to be fixed
- Border
- Assembling of figures inv. under different groups
18Conclusion
- Decomposition using transformation groups is more
powerful - Bottom-up algorithm computing remaining system,
G-reference, border - Principles valid in 3D
19Construction vs Resolution
- Construction
- yield several solutions
- solution space is scanable
- Resolution
- deal with all types of constraint
- numerical solving
20Second example
fixed direction