| title | DeliveryAI |
|---|---|
| emoji | 🚚 |
| colorFrom | blue |
| colorTo | green |
| sdk | docker |
| pinned | false |
SmartRoute is a real-world OpenEnv environment simulating last-mile delivery optimization.
Inspired by systems used by Amazon, Swiggy, and Zomato, this environment enables AI agents to:
- Assign deliveries to vehicles
- Optimize routes under traffic constraints
- Minimize cost, delay, and emissions
Logistics optimization is critical for:
- Faster deliveries
- Reduced fuel consumption
- Lower carbon emissions
This environment models real challenges faced in urban delivery systems.
- Vehicles: location, fuel
- Orders: pickup, drop, priority, deadline
- Traffic conditions
- Current time step
- Assign order to vehicle
- Multi-agent coordination
- Single vehicle
- Static orders
- Goal: maximize delivery success
- Multiple vehicles
- Dynamic traffic
- Deadlines + traffic + coordination
- Goal: optimize efficiency & emissions
- ✅ On-time delivery: +50 × priority
- ❌ Late delivery: penalty
- ⛽ Fuel/traffic cost: negative
- 🌱 Emissions penalty
- 🤝 Coordination bonus
- Delivery success rate
- Total reward
- Efficiency score
- Emissions
pip install -r requirements.txt
docker build -t slda .
docker run -p 7860:7860 slda