Controlling food web structure by optimization of a community assembly model

In community assembly models, species are introduced from a pool of species according to a random sequence of invasion. The present work describes a new approach based on genetic algorithms (GA) which generates non-random sequences in order to maximize the diversity of the community. The GA must also meet the constraint that the food web of the community constructed in this fashion have a specified connectance. We show that the optimized sequences produce communities with a higher diversity than those generated from random assembly sequences. In addition, the GA is able to generate sequences that produce food webs from identical regional species pools that have different expected connectances. The results demonstrate the effectiveness of genetic algorithms for optimizing parameters in ecological models.