An Optimal Group Decision-Making Approach for Cyber Security Using Improved Selection-Drift Dynamics

In cyber security, balancing investment and risk has always been a dilemmatic problem since the threats often lurk in the shadows. Thus, timely and scientifically credible decision-making will have great significance in such an attack-defense game of incomplete information. To this end, an evolutionary game model of group decision-making is proposed to analyze the behavioral change process of the defender population. In particular, we first introduce the concept of observation error and short-term prediction in network communication and establish an improved selection-drift dynamics in which errors are automatically corrected to a certain extent and the convergence speed is faster than the prototype. By calculating the stable evolutionary equilibrium of the defender population, the optimal group decision-making approach is formulated. Case studies on big data vulnerabilities indicate that the proposed model and approach perform better in robustness and computational efficiency than the 3 typical models under jamming environments.