Cyber range environments are widely used for cybersecurity training and experimentation, yet many rely on simplified background user activity that fails to reproduce the ambiguity and operational noise encountered by Security Operations Centers (SOCs). This paper proposes a role- and competence-aware multi-agent user behavior modeling framework for SOC-oriented cyber range evaluation, treating legitimate user activity as a controllable experimental variable rather than passive background traffic. The framework supports structured comparison between simplified and heterogeneous user populations and defines SOC-relevant evaluation metrics, including false positive rate, alert precision, mean time to detect, alert-per-user ratio, and the Analyst Load Index, to assess detection engineering effectiveness and analyst-facing operational burden under realistic behavioral variability. A proof-of-concept simulation with 40 agents shows that role- and competence-aware behavior reduces alert precision from 0.90 to 0.14 and nearly triples the Analyst Load Index while detection coverage remains unchanged, capturing operationally relevant effects invisible to binary detection testing.
- Home
- Publications
- Role- and Competence-Aware Multi-Agent User Behavior Modeling for SOC-Oriented Cyber Range Evaluation