Nonproliferation informatics: Employing Bayesian analysis, agent based modeling, and information theory for dynamic proliferation pathway studies

Decision making on weapons of mass effect (WME) proliferation and counter-proliferation is information driven. However, the large data requirements, along with associated knowledge gaps and intelligence uncertainties, impedes optimal strategy selection. Combining Bayesian analysis, agent based modeling (ABM), and information theory within a security informatics context can aid understanding of dynamic WME proliferation and counter-proliferation pathways and possibilities. The Bayesian ABM Nonproliferation Enterprise (BANE) was developed to incorporate large databases and information sets. There are three broad BANE agent classes: 1) proliferator, 2) defensive, and 3) neutral. Within each agent class exists significant flexibility for them pursuing different objectives. Bayesian analysis cover the technical linkages realistically tying proliferation pathway process steps together. In BANE, Bayesian networks using the Netica software program provide a wide array of scientific and engineering pathway options. Information theory, especially entropy reduction and mutual information, in a Bayesian security informatics arrangement help identify optimal technical areas to master or disrupt. Concurrently, interlocking factors such as available resources, technical sophistication, time horizons, detection risks, and agent affinities impact agents' ability to achieve their goals. Actions taken by one BANE agent on the proliferation or counter-proliferation front affect its future opportunities and those of potential partner or adversarial agents. An explanation of the BANE framework and several key security informatics aspects crucial to WME proliferation and counter-proliferation analysis are provided.