Traditional logistics operations are highly reactive. We provide a comprehensive Admin Command Center that transforms reactive logistics into a proactive, autonomous ecosystem.
The Solution: WareFlow uses Reinforcement Learning for internal warehouse fulfillment and intelligent shortest-path routing between cities.
Unique Selling Proposition (USP): An end-to-end "intelligent" journey where an ML model predicts route disruptions, including both extreme weather and live traffic, to autonomously reroute shipments in real-time.
- Admin Dashboard: A centralized web interface for total oversight of orders, inventory, and global fleet status.
- Autonomous RL Fulfillment: A trained model navigates robots through warehouse grids to pick items with zero human intervention.
- 3D Digital Twin: Live visualization of the warehouse floor directly within the admin view.
- Predictive Risk Engine: Monitors live weather and traffic data to assign risk scores to every transit route.
- Dynamic Rerouting: Automatically recalculates and updates paths in the system when high risks are detected.
- Gemini AI Alerts: Generates natural language reasoning to explain disruption risks to the admin.
WareFlow leverages a suite of powerful algorithms to drive its autonomous operations:
- XGBoost: Employed to efficiently and intelligently select the optimal warehouse based on real-time factors like inventory levels, distance, and fulfillment load.
- Dijkstra’s Algorithm: Implemented for calculating the global shortest path across the 22-city highway graph network.
- Reinforcement Learning (PPO): Powers the autonomous agent's micro-navigation, pathfinding, and obstacle avoidance within the warehouse grid.
- LightGBM: A Gradient Boosting model utilized for high-accuracy prediction of transit risks caused by extreme weather and traffic events.
- Frontend (Admin Command Center): React.js, Three.js, React Three Fiber, Tailwind CSS.
- Backend Ecosystem: Dual-server architecture using FastAPI (Python) for high-performance async processing and WebSockets for live streaming.
- Cloud & Database: Firebase Realtime Database for global state synchronization and Google Cloud Platform for scalable deployment.
- Python 3.10+
- Node.js 20.19+
- A
serviceAccountKey.jsonplaced in thewareflow_p1/directory. - A
backend/.envfile containingGOOGLE_MAPS_API_KEYandGEMINI_API_KEY.
cd backend
pip install -r requirements.txt
python run_backend.pycd frontend
npm install
npm run devNavigate to http://localhost:5173 to access the Admin Command Center.
- Demo Video: Watch here