基于多目标优化的Tor陷阱节点部署策略研究OA

To address the lack of systematic optimization strategies for trap relay deployment in the Tor net-work,this paper proposes a deployment framework utilizing multi-objective optimization and randomized ob-fuscation to enhance both de-anonymization efficiency and node stealthiness.Specifically,the paper,first con-structs a local connectivity graph using the Tor control protocol,and integrates multiple centrality metrics,in-cluding degree,closeness,and Katz centrality.It then quantifies node monitoring visibility by applying min-max normalization and temporal weighting.Subsequently,the trap relay deployment problem is formulated as a multi-objective optimization model aimed at maximizing surveillance coverage,minimizing operational costs,and reducing detection risks.The NSGA-Ⅱ algorithm is then employed to generate Pareto-optimal so-lutions.Concurrently,a Poisson-distribution-based randomized obfuscation mechanism is introduced to dy-namically adjust trap relay activation rates,thereby making their behavioral characteristics resemble those of legitimate relay nodes.Simulation experiments based on real Tor network topology and traffic data demon-strate that,under identical injection scales,the proposed approach significantly reduces the proportion of trap relays flagged as suspicious by SybilHunter while maintaining high monitoring visibility and traffic coverage.Compared with baseline strategies,this framework achieves an effective balance between stealthiness and re-source utilization,demonstrating strong robustness and scalability.The results indicate that combining multi-objective optimization with dynamic obfuscation mechanisms can significantly enhance the survivability and de-anonymization potential of trap relays,providing a quantifiable deployment framework for anonymity sys-tem security evaluation and anonymity research.