Percolation-Based Cloud Computing Framework for Network Reliability Assessment

Effective real-time traffic management is essential for mitigating congestion and ensuring smooth mobility in urban environments. This paper introduces a state-of-the-art cloud computing framework for network trajectory discretization and bottleneck identification using percolation theory for traffic management. The approach evaluates link criticality within transportation networks by integrating traffic flow theory with network science models. Using a taxi trajectory dataset from Daejeon, South Korea, we develop a framework to quickly identify bottlenecks based on link velocity patterns and their impact on network connectivity. A percolation process is applied to progressively remove low-quality links, simulating network degradation and highlighting critical bottlenecks that disrupt origin-destination flows. The framework leverages parallel processing on cloud platforms to expedite the preprocessing of trajectory data and compute link quality and criticality scores in minimal time. This scalability ensures the framework can handle large datasets efficiently, offering traffic management systems dynamic, actionable insights for congestion mitigation. By identifying critical links that significantly affect network reliability, the approach enables targeted interventions for improving traffic flow. This interdisciplinary approach demonstrates the power of cloud-based big data analytics in addressing complex problems in transportation engineering, filling a critical gap in the application of computer science to intelligent transportation systems.