Cellular Evolution in a 3D Artificial Chemistry
The project aims to demonstrate, for the first time, the spontaneous formation and evolution of cells from a "primordial soup", using a simplified, three-dimensional "artificial chemistry". The two-dimensional case taxes the fastest desktop machines available; the three-dimensional case needs hundreds of times more processing power to be feasible.
|
Principal Investigator Duraid MadinaHPC Support Unit University of NSW |
Project f14 |
|
Co-Investigators Naoaki OnoHuman Information Science ATR Kyoto, Japan Takashi Ikegami Graduate School of Arts and Sciences Tokyo University, Japan |
RFCD Codes 270106, 230113, 230101, 280406 |
Significant Achievements, Anticipated Outcomes and Future Work
We were successful. We could, in the end, demonstrate cellular formation and dynamics in an artificial chemistry in three dimensions. The code and initial testing was done on our workstations but a few hundred CPU hours were burned in pursuit of our first "succesful" result. After that, exploratory runs were done, primarily to give the data illustrated in the last figure of the publication mentioned below, consuming the rest of this project's time. In the future, more experimentation with the model will happen, but already some of us are considering a different model, attempting to reproduce the same behaviour but in a more realistic molecular dynamics setting.
Computational Techniques Used
We used stochastic dynamics on a 3D lattice to model a set of reacting, diffusive particles. This model has some fairly substantial changes which distinguish it from typical "reaction/diffusion" models. One version of the code uses the "graphcode" and "classdesc" packages developed as part of the CT&T program. Another is just "straight C + MPI".
Publications, Awards and External Funding
External Funding and Awards
None.
Publications
D. Madina, N. Ono, T. Ikegami, Cellular Evolution in a 3D Lattice Artificial Chemistry, Proceedings of ECAL 2003 (eds. Banzhaf et al, Springer LNCS, 2003, 59-68