Modeling and monitoring of service resilience and reliability for wireless cellular networks through a probabilistic geometrical coverage model
提出概率几何覆盖模型（PGCM）解决不规则蜂窝网络的多重覆盖评估难题，结合比例风险模型和地理统计外推法，在保护隐私前提下监测基站可靠性和用户需求时空场，实现大规模网络韧性与可靠性评估。
The increasing reliance on digital infrastructure in the 5G era necessitates highly reliable wireless cellular networks. To this end, modeling the complex network resilience and reliability is crucial, especially in the face of temporal stochasticity and spatial heterogeneity. A primary challenge in evaluating network resilience arises from the intricate multi-fold geometry associated with irregularly deployed cellular networks, defined as the Multi-fold Coverage Evaluation (MCE) problem. This challenge results in the Challenges of Network Geometry (CNG) using existing formulation and modeling approaches. To address this issue, we propose a novel method called the Probabilistic Geometrical Coverage Model (PGCM). The PGCM employs advanced computational geometry and stochastic field techniques to effectively tackle the MCE problem and overcome the difficulties associated with CNG. To facilitate practical implementation, we introduce a proportional hazard model using a stochastic sequences method. The method allows us to track the reliability of base stations using daily alarm streams. Furthermore, to ensure compliance with privacy regulations, a geostatic extrapolation is implemented to estimate the spatiotemporal field of user demands. It helps in assessing service resilience and reliability, capturing the dynamics of heterogeneous regions even during disruptive events. Through testing on real-world wireless cellular networks, our approach demonstrates its effectiveness in enabling timely evaluations of large networks. Our approach underscores the critical role of resilience in achieving high service reliability and provides valuable insights for privacy-respecting network performance modeling and monitoring.