barnabemonnot / control

Analysis of networks centered on link weights

Weight Analysis

Uses the ZEN Python Library to work on the graphs. The reducer method works on a graph from which it removes a defined number of edge at each step. These edges may be removed depending on their link weights (either increasing or decreasing) or randomly. Several statistics are returned. avg_reducer repeats the procedure a defined number of times to remove the randomness from the choice of edges that possess the same weight.

We also give a few helpful functions to compute statistics on the graph, such as the mean weight from the incident edges of nodes of a certain degree. Helper functions are written to plot graphs illustrating the evolution of controls.