Dia: Dimecres 6 de setembre de 2006.

Lloc: Aula T2 (2on pis), Facultat de Matemàtiques, UB.



Dia: Dimecres 8 de novembre de 2006.

Lloc: Aula T2 (2on pis), Facultat de Matemàtiques, UB.



Dia: Dimecres 29 de novembre de 2006.

Lloc: Aula T2 (2on pis), Facultat de Matemàtiques, UB.



Dia: Dimecres 20 de desembre de 2006.

Lloc: Aula T2 (2on pis), Facultat de Matemàtiques, UB.



Dia: Dimecres 24 de gener de 2007.

A Swinging Atwood's Machine and Non-Integrability day

Lloc: Aula B1 (planta baixa, costat Gran Via), Facultat de Matemàtiques, UB.


Resums:

  1. This presentation focuses on some experimental results relative to the Swinging Atwood's Machine (SAM) dynamics, which, to our knowledge, has not been yet investigated in this way. SAM is a nonlinear system with two degrees of freedom derived from the well-known simple Atwood machine. The latter was built in 1784 by George Atwood - a Physics teacher from London who constructed various apparatus in order to illustrate his Physics lectures - to experimentally demonstrate the uniformly accelerated motion of a falling system in the earth gravity field g with an acceleration lower than g. In Atwood's machine, two masses are mechanically linked by an inextensible thread and a pulley, whereas in SAM one of the mass (m) is allowed to swing in a plane while the other mass plays the role of a counterweight (M), so that SAM is a parametric pendulum with a variable length depending on mu = M/m.

    For about twenty years, many theoretical and numerical studies have been performed about the mechanical behaviour of SAM. Particularly, it has been shown that SAM is a very rich dynamical system exhibiting enormous different trajectories depending on mu-values. For mu > 1 it has been qualitatively shown by using Poincare maps that SAM revealed regular and chaotic behaviours, the latter becoming prominent as mu increases. An interesting and surprising result is for mu = 3 for which SAM appears to be integrable, a conclusion also supported theoretically by Hamilton Jacobi's theory.

    First, we will describe in detail the constructed apparatus by showing some photos of SAM and its different elements. A schematic representation will be then given in order to derive the equations of motion and videos on different experiments performed will be shown; in particular for the value mu = 3. Experimental results coming from analyses of the video sequences will be then presented and a re-analysis of the motion is proposed. Especially the contribution of the pulleys is no more neglected contrary to the previous theoretical studies since experiments show that they can rotate around their revolution axis and that their dimension has to be taken into account. Numerical simulations to solve equations of motion will be also presented and compared with previous studies. Finally, some perspectives for further theoretical studies will be proposed, in particular concerning the integrability of SAM.

  2. Given an autonomous dynamical system with an integral curve Gamma, the variational equations of the system along Gamma are the linear homogeneous system whose principal fundamental matrix is the linear part of the flow of the system along Gamma. A distinguished example, though by far not the only one, is a Hamiltonian system such as the ones commonly studied in Astrophysics. Although such systems are considered real in a widespread number of cases, everything will be considered in the complex analytical setting here.

    So far we have a (generally nonlinear) system and a linear system linked to the former. Heuristics of the non-integrability result central to this talk are firmly rooted in the following: if we assume the initial system "integrable" in some reasonable sense, then the corresponding variational system along any integral curve must be also integrable in the sense of linear Galois differential theory. Any attempt at ad-hoc formulations of this heuristic principle has only one possible drawback: the need for a notion of "integrability" of the original system; but such a specific notion is made available in the Hamiltonian case by the Liouville-Arnold Theorem. With this as a premise, the main result presented in this talk, proven by Juan J. Morales-Ruiz and Jean-Pierre Ramis, is the following: if an n-degree-of-freedom Hamiltonian H has n independent first integrals in pairwise involution defined on a neighborhood of an integral curve Gamma, then, the identity component of the Galois group of the variational equations of X_H along Gamma is a commutative group.

    A general overview on the result will be done, as well as the assessment of some of the basic framework leading, including the preliminary lemmas, one of which is far more than just subservient to the final result. A comment will be made on how a special case, namely that of classical Hamiltonians, allows for a dramatic simplification of the conditions of the main theorem.

  3. Inasmuch as in the previous talk, and although real Hamiltonians are considered for the most part in applications of differential Galois theory, a complex analytical setting is maintained for all Hamiltonian systems X_H considered in the present talk. In such a setting, a conjecture was made by Juan J. Morales-Ruiz calling for a natural extension of the previous Morales-Ramis criterion to higher variational equations; this conjecture was based on suggestion by Carles Simo, in turn stemming from numerical and analytical evidence provided by the study of the higher terms in the jet for a number of special cases. Since, while nonlinear in and of themselves, higher variational equations are easily proven to be equivalent to linear systems (after adjoining an adequate set of redundant variables), Galois differential theory finds no theoretic obstacle to its application in this extended problem.

    With this fact in mind, and in joint work with Ramis, a proof was finally obtained by all three authors for the conjectured result: a necessary condition for the meromorphic complete integrability of X_H is the commutativity of the Galois group G_k of each variational equation of arbitrary order k>0 along any integral curve Gamma.

    With this result rigorously proven, it has subsequently been possible to close important open problems of integrability which had eluded the criterion presented in the previous talk. An example will be exposed here as well as a general outlook on the algebraic framework surrounding the central result.

  4. Abstract not yet available

  5. A preliminary tool is a good integration routine for ODE. High order Taylor methods, with a truncation error small even compared to the computer epsilon, are good candidates. Main integration errors are due to round off errors and the propagation of these ones thourgh the dynamics of the system. With this tool one can face different alternatives to put in evidence the non-integrability:

    1. Poincare maps P: computed on a given level of energy and, eventually, taking also into account other known integrals. First and higher order differentials of P can also be computed numerically.
    2. Computation of Lyapunov exponents: either the maximal one or all of them. Several methods can be used to "filter" the wild behavior and speed up convergence.
    3. Rotation number and, in general, frequency analysis. They are based on topological tools and in a collocation method for the fine tuning of frequencies.
    4. Computation of homoclinic orbits and/or heteroclinic chains and checks of transversality. Using a) this is relatively easy in the case of hyperbolic periodic orbits. It becomes a more subtle problem when the problem is very close to integrable (possible exponentially small phenomena) and in the case of partially normally invariant tori (because of dimensionality).
    5. Direct tests of the failure of the necessary integrability conditions coming from talks (2) and (3).


Dia: Dimecres 31 de gener de 2007.

Lloc: Aula T2 (2on pis), Facultat de Matemàtiques, UB.



ATENCIÓ: S'anuncien dos xerrades. Estretament relacionades amb CAP07


Dia: Dimecres 14 de febrer de 2007.

Lloc: Aula T2 (2on pis), Facultat de Matemàtiques, UB.



Dia: Dimecres 21 de febrer de 2007.

Lloc: Aula T2 (2on pis), Facultat de Matemàtiques, UB.



Dia: Dimecres 28 de febrer de 2007.

Lloc: Aula T2 (2on pis), Facultat de Matemàtiques, UB.



Dia: Dimecres 2 de maig de 2007.

Lloc: Aula T2 (2on pis), Facultat de Matemàtiques, UB.



Dia: Dimecres 9 de maig de 2007.

Lloc: Aula T2 (2on pis), Facultat de Matemàtiques, UB.



Dia: Dimecres 16 de maig de 2007.

Lloc: Aula T2 (2on pis), Facultat de Matemàtiques, UB.



Dia: Dimecres 30 de maig de 2007.

Lloc: Aula T2 (2on pis), Facultat de Matemàtiques, UB.



Dia: Dimecres 13 de juny de 2007.

Lloc: Aula T2 (2on pis), Facultat de Matemàtiques, UB.



Dia: Dimecres 4 de juliol de 2007.

Lloc: Aula T2 (2on pis), Facultat de Matemàtiques, UB.



Dia: Dilluns 9 de juliol de 2007.

Lloc: Aula B1 (planta baixa), Facultat de Matemàtiques, UB.

Hora: 11h

Lectura tesi doctoral de Sergi Simón Estrada

Títol:On the Meromorphic Non-integrability of Some Problems in Celestial Mechanics.


Dia: Dimecres 11 de juliol de 2007.

Lloc: Aula T2 (2on pis), Facultat de Matemàtiques, UB.



Back to the history of seminar

Last updated: Mon Aug 31 12:21:59 MEST 2007