Efficient Allocation of Resources for Defense of Spatially Distributed Networks Using Agent-Based Simulation

Area of Science:

• Operations Research
• Network Security
• Artificial Intelligence

Background:

• Defending critical infrastructure requires distinct strategies compared to natural disaster preparedness.
• Active adversaries anticipate and counter defensive measures, posing unique challenges for resource allocation.

Purpose of the Study:

• To develop an efficient resource allocation strategy for defending spatially distributed physical networks against active adversaries.
• To compare the effectiveness of an agent-based modeling approach with traditional probabilistic risk analysis (PRA).

Main Methods:

• Utilized a combination of integer programming and agent-based modeling.
• Conceptualized the problem as a Stackelberg "leader follower" game.
• Employed deterministic network interdiction and evolutionary agent-based simulation to find evolutionary stable strategies.

Main Results:

• The agent-based approach demonstrated a higher percentage of defender victories.
• Identified evolutionary stable strategies for resource allocation between attackers and defenders.
• Showcased the model's application on an example network.

Conclusions:

• Agent-based modeling offers a superior approach for optimizing defense resource allocation against intelligent adversaries.
• The developed framework provides a more effective strategy than traditional PRA for network defense scenarios.