0_Fleetpy_for_dummies - TUM-VT/FleetPy GitHub Wiki

If you’ve landed here, you probably already know that simulating Mobility-on-Demand (MoD) services is a challenging task. These systems need to capture the complex interactions between various agents—operators, customers, vehicles, etc.—and they rely on advanced fleet control algorithms and diverse data sources.

The figure below illustrates some of the key components that a MoD simulator may need to address:

Fleetpy is here to make your life easier. It’s a flexible, modular, and open-source framework designed to simplify MoD simulation so that you can focus on the specific aspects that matter most to your research— without being distracted by unrelated complexity.

Is Fleetpy for Me? Who Can Use Fleetpy?

Fleetpy is designed to support a wide range of users, regardless of their prior experience with MoD systems. Broadly speaking, there are two main user groups who can benefit the most:

🧠 1. Researchers Focused on MoD Operations & Control Algorithms

Are you working on developing fleet control algorithms—like vehicle repositioning, dynamic dispatching, or request-vehicle matching? Fleetpy is built for that. Thanks to its modular structure, you can plug your algorithm directly into the simulation, either by creating a new module or by extending an existing one. There’s no need to build a full MoD simulation environment from scratch—Fleetpy handles the rest using tested, reliable algorithms.

What does this mean for you?

🔁 You can benchmark your solution against other strategies (some of them already implemented in Fleetpy)
🌐 You can easily test your algorithm in different network settings and with various demand levels.

New Datasets: We recently released network and demand datasets for Munich, Manhattan, and Chicago. They are available in this Zenodo Repository. Example simulation scenarios are available in the Studies folder

🏙️ 2. Planners & Researchers Interested in MoD Impact Analysis

Are you more focused on the broader picture—like how a MoD service would affect transportation in your city? Want to test scenarios with different fleet sizes, vehicle types, or depot locations? Fleetpy has you covered. Simply load your city’s network (e.g., from OpenStreetMap), define the demand and service parameters, and run your simulation. You don’t need to dive into the inner workings of fleet control—Fleetpy comes pre-equipped with state-of-the-art algorithms for routing, passenger-vehicle matching, fleet rebalancing, and more. These are similar to the ones used by real-world operators like Uber and Lyft.

Looking for inspiration?

Check out the Publications section to explore how Fleetpy has already been used.