"Genetically Modified Network Topologies"
We present a mechanism for constructing networks with a given set of parameters using genetic algorithms. The tunable parameters include number of nodes, number of links, clustering coefficient, entropy and average distance. It is shown that the effects of maximizing entropy while constraining the number of links reproduces an exponential degree distribution, as can be seen in many real networks. We also introduce the concept of the Optimal Network Manifold, a boundary in parameter space that constrains a network’s potential characteristics.