Key takeaways
The PhD research will be performed as part of an NWA project funded by the Dutch Research Council (NWO) called "It's a cargo match! - Attaining waste-free and effective freight transport systems by seamlessly matching demand and supply with inclusive, smart, and green-oriented booking platforms". The goal of this project is to conceive a platform of platforms for multimodal freight transport services to match transport jobs with transport resources in dynamic and stochastic environments. Operational matching decisions should be quick and transparent, but at the same time anticipate future uncertainties and strive for high system efficiency. In this context, (multi-agent) reinforcement learning algorithms need to be designed to support online decision-making, both at the individual actor level and at the global system level. Concretely, the PhD candidate will perform research on the following aspects:
- Understand how AI methodologies can facilitate dynamic supply-side decisions
- Understand how AI methodologies can facilitate dynamic demand-side decisions
- Understand how AI methodologies can proactively facilitate dynamic matching of supply and demand
The challenge
The platform economy is on the rise, exemplified by popular booking platforms in the airline- and hospitality industries. These platforms favour synchronization and matching between supply and demand, thereby offering better capacity utilization. In a similar fashion, freight transport services could be matched to transport requests, consequently enhancing sustainability, agility, and resilience of supply chains.
However, deploying such platforms is complex, due to the variety of transport modes and cargo features, conservative business models, and rigid mindsets that limit the opportunities for interconnected networks and visibility of transport options. Thus, there are many hurdles to overcome to design successful platforms for the freight transport sector.
This PhD position focuses on algorithmic design to support online decision-making. By leveraging real-time data, developed algorithms dynamically respond to shifting demand and supply, as well as events such as delays. A combination of approximate mechanism design, market mechanisms, multi-agent system modelling, machine learning, and heuristics procedures will be investigated to support a well-functioning freight transport platform.
Within this project, you will have the opportunity to work not only with colleagues at the HBE department of the University of Twente, but also with researchers from TU Delft, with the opportunity for regular visits to TU Delft, and with our industrial partners from the logistics sector.