Neuro-evolution of escape behaviour under high level of deception and noise

Abstract

Red teaming is an approach to studying a task by anticipating the actions of an adversary ("red") who wishes to affect the achievement (by "blue") of that task. Computational red teaming is a recent approach that extends in red teaming concept in cyber space and benefits from replacing the physical red and blue with simulated entities. In this study, we focus on the use of multiple strategies in computational red teaming and the factors that influence the selection of strategy. The reason for the use of multiple strategies is to simulate variability observed in human choice. The use of multiple strategies are demonstrated by the generation of diversified solutions by evolutionary robotics while the factors that influence the preferences of strategies are perception and deception. This paper presents an attempt at exploring and modeling the effect of red through the evolutionary method in a synthetic red teaming game environment. © 2013 Springer-Verlag.