Inria DRI

International Relations

Internship Subject
PDF Version

2899 - Design and Implementation of a Multi-Market Transport Simulation Framework

This internship aims to design and implement a modular Python-based transport simulator capable of modeling multiple levels of the supply chain — from international maritime cargo to urban last-mile delivery.

 

The simulator will reproduce realistic market settings involving agents (e.g., shippers, carriers, couriers), infrastructure (e.g., ports, hubs, city zones), and operations (e.g., task assignment, routing, scheduling). Each agent must act with its own logic and constraints (time windows, capacity, bundle requirements, etc.).

 

Inspired by open-source tools such as MABLE, FleetPy, or MaaSSim, the intern will first study how such tools handle agent logic and decision structures. Then, they will build their own simulator focused on extensibility, realism, and interfacing with optimization algorithms (e.g., MILP solvers, reinforcement learning agents). This includes the ability to:

 

  • Generate structured decision instances from the simulation (bidding, scheduling, dispatch)
  • Accept allocation decisions from an external solver or API
  • Reintegrate these decisions into the simulated environment dynamically

 

The final simulator should support different market layers (e.g., long-haul shipping vs. last-mile delivery) and enable experimentation with optimization logic across the supply chain.