Scholarlink

<p>This paper proposes a quantitative framework for assessing trust as a dynamic, resilience-relevant property of cyber ecosystems under partial observability. Motivated by increasing reliance on complex, interdependent sociotechnical supply chains, the framework models a cyber ecosystem as a network of interacting nodes whose trustworthiness evolves over time based on internal and external observations. Trust is operationalized using a fading-memory Kalman filter to update node-level trust estimates in response to new information while accounting for uncertainty, subjectivity, and temporal decay. Ecosystem-level trust is then aggregated using a weighted geometric mean to support system-level assessment and comparison. A Monte Carlo analysis evaluates the sensitivity and robustness of the framework across key parameters related to observability, weighting of evidence, and the frequency of disruptive events. Results indicate that internal, judgment-based assessments exert a stronger influence on trust trajectories than the sheer volume of observations, and that increased observability does not necessarily increase aggregate trust, instead revealing latent vulnerabilities within the ecosystem. These findings highlight trust as a diagnostic indicator rather than a measure of confidence alone. The proposed framework supports resilience quantification and benchmarking by enabling decision-makers to track perceived trust dynamics, identify weak points, and prioritize interventions across complex sociotechnical ecosystems. While motivated by cyber supply chains, the approach is broadly applicable to other domains characterized by uncertainty, interdependence, and limited observability.</p>

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

Trust and observability in cyber ecosystems application of a fading-memory Kalman filter and geometric mean to assess dynamic, non-transitive trust in cyber ecosystems

  • 期刊论文

  • 发表日期：

    2026年3月6日

查看全文
  • Ryan P. Hilger,
  • Steve J. Simske

摘要

This paper proposes a quantitative framework for assessing trust as a dynamic, resilience-relevant property of cyber ecosystems under partial observability. Motivated by increasing reliance on complex, interdependent sociotechnical supply chains, the framework models a cyber ecosystem as a network of interacting nodes whose trustworthiness evolves over time based on internal and external observations. Trust is operationalized using a fading-memory Kalman filter to update node-level trust estimates in response to new information while accounting for uncertainty, subjectivity, and temporal decay. Ecosystem-level trust is then aggregated using a weighted geometric mean to support system-level assessment and comparison. A Monte Carlo analysis evaluates the sensitivity and robustness of the framework across key parameters related to observability, weighting of evidence, and the frequency of disruptive events. Results indicate that internal, judgment-based assessments exert a stronger influence on trust trajectories than the sheer volume of observations, and that increased observability does not necessarily increase aggregate trust, instead revealing latent vulnerabilities within the ecosystem. These findings highlight trust as a diagnostic indicator rather than a measure of confidence alone. The proposed framework supports resilience quantification and benchmarking by enabling decision-makers to track perceived trust dynamics, identify weak points, and prioritize interventions across complex sociotechnical ecosystems. While motivated by cyber supply chains, the approach is broadly applicable to other domains characterized by uncertainty, interdependence, and limited observability.