Graph Modeling of Denied Segments for Congestion Detection

Tech Stack: Python, SQL, Pandas, Scikit-learn, NetworkX, SageMaker, AWS, Git

• Devised a novel congestion detection algorithm to optimize efficiency of structured robotic floors:
  • Utilized dynamic graph modeling of denied segment data.
  • Applied subsequent clustering of connected components to identify drive deadlocks.
• Conducted multivariate time series analysis of graph metrics and the floor KPIs:
  • Employed a Vector Autoregressive model for detecting congested drive clusters.
• Developed a Python module for:
  • Real-time congestion detection.
  • Post-analysis diagnosis.
  • Enhancing drive path planning services and improving floor layout design.