Abstract
Deconstructing social networks is a matter of particular importance when considering terror or criminal networks. Apart from discussions of general strategies, identifying which nodes or edges should be removed from a network to pull it apart fastest is a subject on which little research has been done. This thesis proposes a method to deconstruct networks based on network flow algorithms that can identify edges which bottleneck a network. The models test the efficiency of this method against two standard strategies across several ideal network types.