Vehicle Allocation System Based on Traffic Congestion

A machine learning-powered system that optimally allocates vehicles in urban networks based on real-time traffic congestion predictions. This project combines graph theory with predictive modeling to improve traffic flow efficiency and reduce congestion.

🚀 Features

• Traffic Prediction: Uses Random Forest Regressor to predict bus counts based on historical traffic data
• Graph-based Route Analysis: Implements NetworkX for modeling street networks and finding optimal paths
• Dynamic Vehicle Allocation: Allocates vehicles based on predicted congestion and road capacity
• Interactive Input System: User-friendly interface for route planning with date/time specifications
• Visualization: Network graphs showing traffic flow and vehicle allocation

📊 Dataset

The system uses Chicago Traffic Tracker historical congestion estimates by segment from 2018, containing:

• Segment IDs for road identification
• Date and time information
• Bus count data for traffic volume
• Street segment information (FROM_STREET, TO_STREET)
• Road length data for capacity calculations

🛠️ Technologies Used

• Python Libraries:
  • pandas & numpy - Data manipulation and analysis
  • scikit-learn - Machine learning models (Random Forest Regressor)
  • NetworkX - Graph creation and path analysis
  • matplotlib - Data visualization
  • datetime - Time-based calculations

🏗️ System Architecture

1. Data Preprocessing: Historical traffic data is cleaned and prepared
2. Feature Engineering: Date conversion to ordinal format for model training
3. Model Training: Random Forest Regressor learns patterns from historical data
4. Graph Construction: Street network represented as nodes (intersections) and edges (road segments)
5. Path Finding: Shortest path algorithms identify optimal routes
6. Congestion Prediction: ML model predicts traffic for specific date/time
7. Vehicle Allocation: Dynamic allocation based on predicted congestion and road capacity

📈 Model Performance

The system tracks model accuracy using:

• RMSE (Root Mean Square Error): Measures prediction accuracy
• MAE (Mean Absolute Error): Average prediction deviation

🎯 How It Works

1. Input Collection: Users specify origin, destination, date, and time
2. Path Analysis: System finds all shortest paths between locations
3. Traffic Prediction: ML model predicts congestion for each road segment
4. Allocation Calculation: Vehicles allocated based on:
   • Predicted traffic volume
   • Road segment length/capacity
   • Time-based traffic patterns
5. Visualization: Results displayed as network graphs with allocation data

🚦 Real-World Applications

• Public Transit Planning: Optimize bus routes and frequencies
• Emergency Services: Efficient ambulance and fire truck deployment
• Ride-sharing: Dynamic vehicle positioning for services like Uber/Lyft
• City Planning: Data-driven decisions for traffic infrastructure
• Logistics: Delivery vehicle routing optimization

📝 Usage

The system provides an interactive interface where users can:

• Enter starting and destination locations
• Specify travel date (year, month, day)
• Select time slot (24-hour format with hourly intervals)
• View predicted congestion and optimal vehicle allocation

🔮 Future Enhancements

• Real-time data integration
• Multi-modal transportation support
• Weather impact modeling
• Mobile application development
• Integration with city traffic management systems

🤝 Contributing

This project demonstrates the practical application of machine learning in urban transportation systems. The combination of predictive modeling and graph algorithms provides a robust foundation for intelligent traffic management solutions.