Resistência de redes de Barabási-Albert sob ataques intencionais

Abstract

Throughout this work, physical and mathematical techniques were employed to study an algorithm capable of identifying and classifying the most important elements for the connectivity of complex networks, particularly Barabási-Albert networks, which exhibit behaviors such as the existence of hubs, self-organization, degree distributions following power laws, and more. In the end, the efficiency of this method is demonstrated through a direct comparison of how a BA network behaves when subjected to separate attacks using different techniques. This was achieved by exploring concepts of ideal percolation applied to complex networks, leading us to the metric of collective influence. This classification method takes into account the quality of connections rather than just the number of connections a node has. Such an approach allows us to identify elements that would otherwise be overlooked in a superficial analysis when attempting to determine the most influential components of a system. Beyond its functionality in neutralizing or protecting complex networks, collective influence can also be used to study the system itself by analyzing the set of most important nodes according to this tool.