Understanding complexity through simulation

Brussels, 24 Jan 2006

According to Wikipedia, the first large-scale application of computer simulation was during the US-led Manhattan Project in the 1940s, when it was used to model the process of nuclear detonation.

Since then, simulation techniques have developed hand-in-hand with the rapid evolution of the computer to the point where, today, they are not only used as vital tools in cutting-edge fields of science and discovery, but equally contribute to the box office success of the latest blockbuster films.

To communicate the potential of simulation technologies to a broader audience, a major conference on 'The age of simulation' was held recently in Linz, with the support of the Austrian Presidency. One of the event organisers, Harald Katzmair, managing director of FAS.research, spoke to CORDIS News about the role that simulation technologies play some 60 years after their emergence.

'There are many reasons why simulation is important, both in science and business,' began Dr Katzmair. 'For me, simulation helps us to explore and understand complex systems. It's not about predicting the future, but rather trying to understand complexity or the rules that underlie complexity.'

Most complex systems are non-linear, which makes it very difficult to predict future behaviour, explained Dr Katzmair. Instead, there tends to be critical mass points, for example in the popularisation of an idea or the spread of a virus, when the system begins exhibiting new or strange behaviour. 'Through simulation we can see when these critical mass points occur,' he added.

In his own work with FAS.research, Dr Katzmair focuses on network analysis and agent based simulation. 'That's why we are interested in CORDIS - for analysing network patterns of cooperation in Europe,' he said. His group took a set of agents representing individual participants in EU research projects and assigned to them a probability of collaboration with another agent of similar attributes (for example, same country of origin or scientific discipline). By analysing the distribution of probabilities in a simulation that mirrors the 'real' observed data, Dr Katzmair can then learn more about the micro motives that drive the agents.

'We now have a good idea of how a network should be structured in order to result in the diffusion of ideas and knowledge - or innovation,' revealed Dr Katzmair. 'The problem is that you cannot structure networks from the top down - they grow out of micro motives. But the knowledge gained by using simulation to understand what rules at the micro level will affect the structure on a macro level is very important for analysis and evaluation. We can then see what critical mass of agents would be needed to change a network, and begin to discover the interrelationships between the micro and macro levels.'

Indeed, Dr Katzmair and his team have extracted from the CORDIS database all projects whose titles or abstracts refer to simulation, allowing them to build a picture of a virtual European network on simulation.

Another recent and highly successful example of agent-based simulation can be found in the entertainment industry, in computer games and Hollywood films. 'The entertainment industry has developed very complex and advanced methods of applying simulation,' explains Dr Katzmair. He offers the example of large battle scenes involving thousands of agents in films such as The Lord of the Rings: 'There are very complex simulations behind these scenes to obtain the realistic actions of the characters, but the advantage is you only have to programme their behavioural rules, you don't actually have to direct them.'

Other 'hot topics' in simulation include diffusion studies, which analyse how innovations diffuse into the market, and are of obvious interest to economists and businesses. Simulation is also being used to analyse the spread of viruses or bio-contamination, which can be used by governments to shape their response to pandemic outbreaks or terrorist attacks. On this last point, Dr Katzmair reveals: 'Lots is being done in relation to terrorism at the moment, for example, using simulation to analyse the structure of terrorist networks.' By simulating the structure of wider terror networks for which the authorities have only fragmented information at present, they can analyse what impact the removal of certain agents would have on the overall network. 'Based on this information, for example, you can then decide whether it's more important to isolate specialists from the network, or concentrate on the decision makers. The US is funding a lot of work in these areas at the moment,' he added.

Europe is lucky to have developed real expertise in the field of simulation, according to Dr Katmair. 'This is something of a technology for the future - it is really the only way of analysing the behaviour of complex, dynamic systems.' The current move towards more a transdisciplinary approach to simulation simply reflects the fact that many large-scale systems share a common complexity resulting from interconnections between individual components or agents. 'Whether you're trying to simulate economic growth, the behaviour of bees or the spread of fire, the transdisciplinary paradigm is of value to all areas,' concludes Dr Katzmair.

Further information on the event 'The age of simulation'

CORDIS RTD-NEWS/© European Communities, 2005
Item source Previous Item Back to Titles Print Item