DSVRPTW

Dynamic Stochastic Capacitated Vehicle Routing Problem with Time Windows (DSVRPTW)

The DSCVRPTW is a generalization of the classical Capacitated Vehicle Routing Problem with Time Windows, incorporating uncertain and dynamic changes during operation: not all information is known offline, as the environment evolves due to dynamic customer requests that arrive sequentially while vehicles are en route, and stochastic travel times that are random and unpredictable, requiring that all vehicle capacity and customer time window constraints be met despite these dynamic changes (see [BONO2021]).

Here’s everything about DSVRPTW environment:

• Agent Rewards
• Agent Selector
• Environment
• Generation
• Observations

References

[BONO2021]

Bono, G., Dibangoye, J. S., Simonin, O., Matignon, L., & Pereyron, F. (2021). Solving Multi-Agent Routing Problems Using Deep Attention Mechanisms. IEEE Transactions on Intelligent Transportation Systems, 22(12), 7804–7813;