Evolving Gaming Strategies for Attacker-Defender in a Simulated Network Environment

This work investigates an evolutionary approach to generate gaming strategies for the Attacker-Defender or Intruder-Administrator in simulated cyber warfare. Given a network environment, attack graphs are defined in an anticipation game framework to generate action strategies by analyzing (local/global) vulnerabilities and security measures. The proposed approach extends an anticipation game (AG) framework by taking into account multiple conflicting objectives like cost, time, reward and performance for generating effective gaming strategies. A gaming strategy represents a sequence of decision rules that an attacker or the defender can employ to achieve his/her desired goal. In this work, a memory-based multi-objective evolutionary algorithm (MOEA) is implemented in AG framework to generate action strategies, and experiments are performed in a simulated network. Simulations with different types of nodes and services are performed, results are analyzed and reported. These experiments demonstrate that the proposed MOEA approach performs better than existing AG implementations.