Archivio 2012
112
seminari
A Symbolic Circuit Simulator is described and implemented within the Mathematica scientific environment. Four lectures will be given on this subject. The lectures are particularly devoted to students enrolled in the Computational Mathematics course, in the Degrees: Laura Magistrale in Informatica, Laura Magistrale in Matematica, Laurea Magistrale in Scienze di Internet. All interested people, though, are welcome to attend. The lecturer is Dr. Oliver Ruebenkoenig, Research and Development Department, Wolfram Research, Urbana - Champaign, Illinois, USA. Further to be a Mathematica Kernel Developer, Dr. Ruebenkoenig's expertize is in the following research areas: - PDEs in general - Finite Element Method - Symbolic finite elements (for automatic code generation) - Finite Volume, Finite Difference - PDE modeling (using PDEs to model engineering applications, like structural mechanics, fluid flow) - Meshing (computational geometry) - Model order reduction (MOR) - Parametric PDE models - Practical computer science in general / data structures - Lumped system modeling (e.g. circuit simulation) He will stay in Bologna from 12/12/2012 and 19/12/2012. For any information, please refer to my email (giulia.spaletta@unibo.it).
A Symbolic Circuit Simulator is described and implemented within the Mathematica scientific environment. Four lectures will be given on this subject. The lectures are particularly devoted to students enrolled in the Computational Mathematics course, in the Degrees: Laura Magistrale in Informatica, Laura Magistrale in Matematica, Laurea Magistrale in Scienze di Internet. All interested people, though, are welcome to attend. The lecturer is Dr. Oliver Ruebenkoenig, Research and Development Department, Wolfram Research, Urbana - Champaign, Illinois, USA. Further to be a Mathematica Kernel Developer, Dr. Ruebenkoenig's expertize is in the following research areas: - PDEs in general - Finite Element Method - Symbolic finite elements (for automatic code generation) - Finite Volume, Finite Difference - PDE modeling (using PDEs to model engineering applications, like structural mechanics, fluid flow) - Meshing (computational geometry) - Model order reduction (MOR) - Parametric PDE models - Practical computer science in general / data structures - Lumped system modeling (e.g. circuit simulation) He will stay in Bologna from 12/12/2012 and 19/12/2012. For any information, please refer to my email (giulia.spaletta@unibo.it).
A Symbolic Circuit Simulator is described and implemented within the Mathematica scientific environment. Four lectures will take place, all at the Ranzani Laboratory (Via Ranzani 14/b, 1st Floor). The lectures are intended for students of the Computational Mathematics course, within the Degree Courses: Laurea Magistrale in Informatica, Laurea Magistrale in Matematica, Laurea Magistrale in Scienze di Internet. All interested people, though, are very welcome to attend. The lecturer is Dr. Oliver Ruebenkoenig, Kernel Developer, Research and Development Department, Wolfram Research, Urbana - Champaign, Illinois, USA. Further to be a Mathematica Kernel devoper, Dr. Ruebenkoenig is an expert in the following research areas: - PDEs in general - Finite Element Method - Symbolic finite elements (for automatic code generation) - Finite Volume, Finite Difference - PDE modeling (using PDEs to model engineering applications, like structural mechanics, fluid flow) - Meshing (computational geometry) - Model order reduction (MOR) - Parametric PDE models - Practical computer science in general / data structures - Lumped system modeling (e.g. circuit simulation)
Il conferenziere è autore, insieme a Ernst Hairer, del testo "L'analyse au fil de l'histoire", tradotto in inglese col titolo "Analysis by Its History".
Si tratta di un testo per un primo corso di analisi in cui i concetti vengono illustrati insieme alla storia della loro scoperta.
Egli giustifica come segue il suo interesse in questo argomento.
I never liked the dust-dry Bourbaki-style instruction Def.-Thm.-Proof (Germans call it the Landau-style, Greeks may call it the Euclid-style).
After years-long experiences of teaching I got more and more convinced that it is not at all suited for first-year courses, in particular for students whose main direction was physics, computer science or chemistry.
I am not the first with this opinion; emphasizing the history for understanding the development of mathematics, its motivations, its difficulties to overcome, its contacts to neighboring disciplines, was normal in earlier times (for example for Pappus, Wallis, Barrow, Euler, Lacroix, Dirichlet, Felix Klein, Toeplitz).
A Symbolic Circuit Simulator is described and implemented within the Mathematica scientific environment. Four lectures will take place, all at the Ranzani Laboratory (Via Ranzani 14/b, 1st Floor). The lectures are intended for students of the Computational Mathematics course, within the Degree Courses: Laurea Magistrale in Informatica, Laurea Magistrale in Matematica, Laurea Magistrale in Scienze di Internet. All interested people, though, are very welcome to attend.
The lecturer is Dr. Oliver Ruebenkoenig, Kernel Developer, Research and Development Department, Wolfram Research, Urbana - Champaign, Illinois, USA. Further to be a Mathematica Kernel devoper, Dr. Ruebenkoenig is an expert in the following research areas:
- PDEs in general
- Finite Element Method
- Symbolic finite elements (for automatic code generation)
- Finite Volume, Finite Difference
- PDE modeling (using PDEs to model engineering applications, like structural mechanics, fluid flow)
- Meshing (computational geometry)
- Model order reduction (MOR)
- Parametric PDE models
- Practical computer science in general / data structures
- Lumped system modeling (e.g. circuit simulation)
A polytope is the convex hull of finitely many points in R^d. Given two arbitrary vertices of a d-dimensional polytope with n facets, how far away can they be? That is, how many edges do we have to walk on, to go from one to the other? The question comes from optimization, as a worst-case scenario from the simplex algorithm. Let us say that a d-dimensional simplicial complex with n vertices is "Hirsch" if its dual graph has diameter smaller than n-d. The Hirsch conjecture (1957) guessed that the boundary of every (d+1)-polytope is Hirsch. The conjecture has recently been disproved by Paco Santos (Annals of Math., 2012). So the bound n-d is wrong; but it could be that 2n is the correct guess; or maybe 2n is also wrong, but dn is the correct one... We really don't know much: At the moment we cannot even prove a *polynomial* upper bound in n and d.
We will present some recent progress (joint work with Karim Adiprasito): The Hirsch conjecture holds true for flag polytopes, and more generally, even for flag cohen-Macaulay complexes. In particular, the barycentric subdivision of every triangulated manifold is Hirsch.
The proof uses a metric geometry criterion by Gromov, but is otherwise elementary; I will start from defining what a simplicial complex is.
We consider the numerical approximation of elastic problems for incompressible materials, in the framework of the large deformation regime. In particular, a number of Galerkin schemes are investigated, ranging from displacement-based finite elements to mixed finite elements and NURBS-based approximations. Our focus is mainly concerned with the capability of the numerical methods under consideration to appropriately detect bifurcations and/or limit points.
To this aim, we propose a couple of simple problems, for which some theoretical results about the stability range are available. We then show that several schemes, efficient and reliable in the infinitesimal elasticity regime, may fail in reproducing the stability behaviour in the large deformation context. We recognise that this failure has its root in the relaxation of the incompressibility constraint, that many methods need to introduce in order to avoid volumetric locking effects. Some numerical results are presented to confirm the theoretical considerations.
This work has been developed in collaboration with F. Auricchio, L. Beirao da Veiga, A. Reali, R.L. Taylor and P. Wriggers.