Why simulating evolutionary processes is just as interesting as applying them

Katharina Lehmann

Evolutionary algorithms are very efficient tools to find a near-optimum
solution in many cases. Until now they have been mostly used to find results but
in this article we argue that evolutionary algorithms can also be used
to simulate the evolution of complex systems. We model
complex systems as networks in which agents are connected by edges if they
interact with each other. It is known that many networks of this kind 
exhibit stable properties despite the dynamic processes they are subject to.
We show here how evolutionary processes on complex systems can be modeled with a new kind 
of evolutionary algorithm which we have presented in \cite{lk-eaftsoeon-05}. 
We will show that some evolutionary processes within this framework yield
networks with stable properties in reasonable time. An understanding of
what kind of evolutionary processes will produce what kind of network properties in what time
is vital to transfer  evolutionary processes to technical ad-hoc
networks in order to improve their flexibility and stability in quickly changing environments.