Supercomputers to simulate life on earth

Switzerland's Brutus supercomputer and several others at the Los Alamos National Laboratory in the U.S. are part of a scientific attempt to simulate life on earth.

In an effort led by the Swiss Federal Institute of Technology, also known as ETH Zurich, researchers hope to explore issues like economic crises and how international conflict originates.

The project is dubbed FuturIcT and supported by 1 billion EUR (A$1.46 billion) in funding from the European Commission's Flagship Programme over ten years.

Researchers say that the project cost is dwarfed by losses from financial crises, which it hopes to avoid using what they call 'socio-physics'.

ETH sociologist Dirk Helbing hopes to apply complex systems theory, statistical physics and evolutionary game theory to describe the interactions of people in social networks, space and time.

"IT has brought the progress that makes large-scale, integrated modelling of socio-economic systems possible," he told iTnews.

"This concerns the availability of computer power needed to simulate complex systems, of tools to model them, and of data to validate and calibrate the models."

Helbing would not provide an estimate of how much supercomputing muscle would be required, noting that it would depend on how "fine-grained" the researchers' models were.

"One can generally say that models become more systemic/holistic on the one hand and more fine-grained on the other hand as more supercomputer power becomes available," he said.

The researchers plan to augments their simulations with social, economic and environmental data from the Internet and massive multiplayer online (MMO) worlds, as well as information from distributed sensor networks.

Data mining would be overseen by an ethics committee, ETH states, noting that the goal is to "identify statistical interdependencies when many people interact, but not to track or predict individual behaviour."

Helbing told iTnews that games like Second Life and The Sims could shed light on some aspects of society, and be a sandbox for alternative system designs before they are implemented in the real world.

However, he said more research was needed into how well in-game interactions translated to reality, and FuturIcT simulations would be rather more complex.

"Complex systems can be best described by a pluralistic modelling approach ... It will certainly not be something like a world formula or a theory of everything," he told iTnews.