Research at the Department for Analysis of Partial Differential
Equations at the Johann Radon Institute for Computational and Applied
Mathematics (RICAM)
Extract from the research site of RICAM
This research group will work mainly on modelling with
and analysis of partial differential equations. It will provide a direct
interface to the groups of the Radon Institute in financial mathematics
(Schachermayer) and of Engl/Langer in inverse problems/numerical analysis
and simulation. The interaction will be very tight since modern pde
modelling/analysis and numerics/simulation cannot survive without one
another in the near future. Although we fully appreciate the fact that
research guidelines have to be extremely flexible (also pushed by an
unprecedented hardware development which nowadays makes simulations
possible which nobody dreamt about only ten years ago), some main new
topics, which will be developed within the Radon Institute, have
emerged.
1) The 'next' generation of pde-analysts will to a large
extent be concerned with non-linear system of pdes. It will be a
main challenge to overcome already well-developed tools for scalar
problems (like maximum principles) pushing the so far underdeveloped
theory of pde systems. 2)Continuing research on (fully) non-linear
pdes: Monge-Ampere equation, Monge-Kantorovich mass transportation
problems, Wasserstein metric, entropies for diffusive systems, convex
Sobolev inequalities, Hamilton-Jacobi equations, free boundary problems
(typical technique: viscosity solutions), connections to control theory,
geometric motions (mean curvature flow, pattern formation),
Hamiltonian-Lagrangian (quantum) mechanics, homogenisation of non-linear
pdes. There are close links to level set methods as mentioned in the
description of the Inverse Problems Group. 3) stochastic pdes with
space-time stochasticity (random Schrödinger equations, Brownian motions
on manifolds, random diffusion systems, stochastic non-linear pdes).
Typical applications are in quantum mechanics (impurity scattering),
geophysics, finance mathematics, mathematical biology. Already ongoing
research of the group of Peter Markowich on models for the dynamics of
large particle ensembles will continue. This covers the range from basic
quantum mechanical models (e.g., (non-linear) Schrödinger and Dirac
equations) to classical macroscopic fluid models, including
(semi-)classical and quantum kinetic models (e.g., Boltzmann,
Fokker-Planck, Wigner equations). A particular emphasis will be on scaling
limits and the analysis of large time behaviour. Specific topics
are: 4) Kinetic models in population dynamics and mathematical biology:
the dynamic interaction of genotype/phenotype and spatial (geography)
effects will be studied by means of kinetic modelling (the
"genotype/phenotype" variables play the role of the momentum variable,
dual to the spatial variable). Chemotaxis will be studied on the kinetic
level, with particular emphasis ondiffusion limits and microscopic
modelling of the motion of bacteria. It is conceivable that finite-time
blow-up can be avoided on the kinetic level by careful modelling of
reorientation processes. 5) Macroscopic limits of kinetic models:
Applications include charge transport in semiconductors, ionisation of
gases, and kinetic models for wave-particle scattering. An understanding
of fluid limits of the latter might lead to a new approach for
understanding macroscopic models of gas dynamics. 6) Modelling and
Simulation of Bose-Einstein condensation: Bose-Einstein condensation is
one of the extremely hot topics in low-temperature atomic physics (cf. the
Nobel Prize in physics 2001). The typical model after condensation is the
Gross-Pitaevskii equation (cubically non-linear Schrödinger equation with
harmonic confinement), while the bosonic Boltzmann equation is used to
describe the condensation process. We expect to continue numerical
simulations and the analysis of the condensation process (transition
regime of the bosonic Boltzmann equation to the Gross-Pitaevskii
equation). 7) Derivation, analysis and numerics of quantum Boltzmann
equations by means of Wigner transforms: This work is supposed to shed new
light on the origin of irreversibility in semi-classical and other scaling
limits. New insights into open quantum systems are expected by employing
quantum entropy and quantum entropy-dissipation techniques similarly to
the classical-mechanics case. 8) inverse problems in fluid-type pdes:
applications in semiconductors, nano-technology, geophysics. In this area
there will be significant interaction with semiconductor device engineers
outside the Radon Institute and with the in-house groups of Engl and
Langer. Main topics will be doping profile identification and device
performance
optimisation.
|