HauensteinJD042009D.pdf (298.71 kB)
Regeneration, local dimension, and applications in numerical algebraic geometry
Algorithms in the field of numerical algebraic geometry provide numerical methods for computing and manipulating solution sets of polynomial systems. One of the main algorithms in this field is the computation of the numerical irreducible decomposition. This algorithm has three main parts: computing a witness superset, filtering out the junk points to create a witness set, and decomposing the witness set into irreducible components. New and efficient algorithms are presented in this thesis to address the first two parts, namely regeneration and a local dimension test. Regeneration is an equation-by-equation solving method that can be used to efficiently compute a witness superset for a polynomial system. The local dimension test algorithm presented in this thesis is a numerical-symbolic method that can be used to compute the local dimension at an approximated solution to a polynomial system. This test is used to create an efficient algorithm that filters out the junk points. The algorithms presented in this thesis are applied to problems arising in kinematics and partial differential equations.
History
Date Modified
2017-06-02Defense Date
2009-03-27Research Director(s)
Andrew J. SommeseCommittee Members
Charles W. Wampler Bei Hu Chris PetersonDegree
- Doctor of Philosophy
Degree Level
- Doctoral Dissertation
Language
- English
Alternate Identifier
etd-04142009-101147Publisher
University of Notre DameProgram Name
- Mathematics
Usage metrics
Categories
No categories selectedLicence
Exports
RefWorks
BibTeX
Ref. manager
Endnote
DataCite
NLM
DC