Advanced School and Workshop on Computational Gravitational Dynamics
Organizers: Alice Quillen, Steve McMillan, Simon Portegies Zwart, Joachim Stadel
Topics: planetary dynamics, star cluster dynamics, galaxy dynamics and cosmology.
The advanced school and workshop on computational gravitational dynamics was attended by some 24 young apprentice researchers and an some 32 experienced researchers world wide.
Computational astrophysics is a fast growing discipline which envelopes computational science, astronomy and physics. There is a range of expert fields within computational astrophysics, some of which seem obviously related but are socially and professionally well separated. We will organize an advanced school and workshop to bridge gaps between a range of specialized fields in computational astrophysics, in particular on those disciplines that simulate gravitational dynamics using $N$-body techniques.
During the advanced school grouped the students to work in collaboration with a experienced researcher (in parenthasis) on one of the following topics:
Orbital Dynamics of Multi-Planet Systems w/GPUs (Prof. E Ford)
Planetary system stability (Prof. A. Quillen)
Toomre & Toomre Redux (Prof. J. Dubinski)
Galactic disk and bar formation (Prof. Ch. Boily)
The Final Parsec Problem (Dr. S. Harfst)
Exploring galaxy collision products (Dr. I. Pelupessi)
Forming ultra-compact dwarf galaxies (Dr. H. Baungardt)
Do spiral galaxies contain a dark matter disc? (Dr. P. Teuben)
Collisional dynamics with a tree code (Dr. J. Stadel)
After this first week, we continued with a more classical approach of a workshop, for which we had invited some 24 experienced researchers. During the workshop the student-teams were encouraged to present their results to the more experienced audience.
We achieved our prime objective in cross fertilizing the computational science aspects of the four core-disciplines in computational gravitational dynamics: planetary dynamics, star cluster dynamics, galaxy dynamics and cosmology. New research project have been initiated during the school. Among these are a collaborative effort to model a real star cluster using a Barnes-Hut tree-code, a tool often adopted by cosmologists and galaxy modelers.
We have achieved in bringing a broad range of students and researchers from domains which are classically well separated. We hope to see the researchers in the 'other' domains more frequently at our local domain-specific conferences, as we have layed the seeds for such further cross fertilization during the school and workshop.
We made important progress in initiating cross disciplinary software tools. It turns out that some of the domain-specific application algorithms are rather specialized and hard to port across the disciplines. But several of our methods appear to transfer quite easily across domains, such as planetary system evolution in context of cluster dynamics, or improving orbit approximations inside other than Keplerian potential to improve accuracy of tree or direct codes other potentials). However, we did not settle on a specific fixed approach that was a clear winner. The workshop was pretty intense and covered a lot of ground so it may take some time for the pieces and new collaborations to come clearer. But surely the seeds for new ideas/approaches were planted, and in due time we expect a flurry of progress in the interdisciplinary research field of computational gravitational dynamics.
We think that it was a good idea to limit ourselves to computational gravitational dynamics. During the school and in particular during the workshop we already had some difficulty understanding the other experts from a different specific research area. Broadening-up the domain could easily result in a Babylonian situation; where most of the time would be spend in trying to underhand each other without direct progress.
We are very enthusiastic about the combination of an advanced school and a workshop, as it gain the interest of a new generation of young students and establish cross fertilization among the different physical scales which utilize similar numerical techniques. A next proposal to the Lorentz Center will certainly have these same two ingredients.