Link aggregation process for modelling weighted mutualistic networks

Abstract

Mutualism is a biological interaction mutually beneficial for both species involved, such as the interaction between plants and their pollinators. Real mutualistic communities can be understood as weighted bipartite networks and they present a nested structure and truncated power law degree and strength distributions. We present a novel link aggregation model that works on a strength-preferential attachment rule based on the Individual Neutrality hypothesis. The model generates mutualistic networks with emergent nestedness and truncated distributions. We provide some analytical results and compare the simulated and empirical network topology. Upon further improving the shape of the distributions, we have also studied the role of forbidden interactions on the model and found that the inclusion of forbidden links does not prevent for the appearance of super-generalist species. A Python script with the model algorithms is available.