Genetic algorithm/artificial life evolution of security vulnerability agents

Although tactical battlefield networking environments of the 21 ^st century are viewed as providing greatly enhanced capabilities compared to today´s environments they are still characterized as unstable (in terms of resource demands, network topology, and network membership), mobile, heterogeneous, resource constrained and at uneven risk. Assessing the resources of such an environment for security vulnerabilities presents a number of unique challenges toward achieving effective and efficient operation. We describe preliminary results of research aimed at addressing vulnerability assessment of tactical battlefield environments through the application of the genetic algorithm and artificial life techniques. Our most critical departures from the “pure” algorithm approaches are (1) the complete loss of a genetic type is not allowed, and (2) that only a fraction of the population is allowed an interaction with the environment at any given time