The project Algorithms for Road Transport Decision Support, AI4ROAD, aims to develop advanced planning solutions by integrating traditional operations research methods with artificial intelligence techniques. Our goal is to create tools that make road transport more efficient, resilient, and sustainable. These solutions address key challenges in the sector, including driver shortages, the ongoing automation of logistics, and the evolving dynamics of collaboration and competition within the transportation market.

In close collaboration with our industry partner, we focus on developing planning tools that are not only realistic and effective but also easy to use, especially for mid-sized transport companies. At the heart of every algorithm and planning tool lies a strong commitment to sustainability, such as reducing empty mileage and improving vehicle capacity utilization.

Industry Partner: Van der Wal

Van der Wal is an international logistics service provider headquartered in Utrecht, employing over 400 people. The company offers a wide range of logistics services, including transportation, flexible storage, and supply chain optimization. It operates as a 2PL (carrier), 3PL (freight forwarder), and 4PL (logistics orchestrator), managing complete transport operations for its clients.
Van der Wal is dedicated to delivering the most sustainable transport solutions possible.

Targeted Areas

1. Next-Day Planning Algorithms
Transport companies must coordinate vehicles, loads, and drivers; each constrained by complex regulations (e.g., country-specific vehicle permits, driver working hours). We aim to develop efficient next-day planning algorithms to optimize these interdependent resources while ensuring regulatory compliance.

2. Dynamic Real-Time Planning
Beyond static planning, we develop capabilities for real-time decision-making. By combining classical optimization techniques with reinforcement learning, we support dynamic planning of trucks, drivers, and shipments in response to disruptions or newly available information.

3. Learning the Planning Logic
To better support human planners, we analyze their decision-making processes using historical data and inverse optimization techniques. This helps reveal implicit constraints, practical heuristics, and anticipatory decisions that shape real-world planning. We aim to build tools that naturally support and enhance planners’ decision-making workflows.

Team members

Casper Bazelmans, PhD student
Neslihan Cevik, PhD student
Dr. Albert Schrotenboer
Prof.dr. Tom Van Woensel

Publications

Logistiek.nl July 2025: Van der Wal en TU Eindhoven proberen planpuzzel te kraken met AI – VIDEO