AI Model Prefers Efficient Grocery Delivery
rotterdam, donderdag, 21 augustus 2025.
PhD candidate Liana van der Hagen from Erasmus University has developed an AI model that determines within milliseconds whether a delivery request can still fit into the schedule. By training the system with thousands of customer request simulations, it can quickly and accurately predict whether a delivery time slot is feasible, leading to better customer satisfaction and more efficient logistics.
Efficient Delivery Planning through AI
The AI model developed by PhD candidate Liana van der Hagen is trained with thousands of customer request simulations. This makes it possible to determine within milliseconds whether a new delivery request can still fit into the schedule. Traditional methods can take hours, while customers want to know immediately which delivery time window is available. According to Van der Hagen: ‘The customer does not want to wait; they want to know right now which delivery time window is available.’ [1]
Collaboration and Future Prospects
The development of the AI model is a collaboration between Erasmus University, Albert Heijn, and ORTEC within the CILOLAB project. The model is currently a proof of concept and has not yet been implemented in software systems. However, Van der Hagen emphasises its significant potential: ‘It is a proof of concept that demonstrates how machine learning can make such decisions much faster than traditional methods used in practice.’ [1] The next step is to incorporate delivery costs into the predictions, guiding customers towards ‘green time slots’. This would lead to fuller delivery vans and a more sustainable and profitable process. [1]
Practical Example: Albert Heijn
Albert Heijn and PhD candidate Liana van der Hagen have collaborated on developing a model that predicts within a fraction of a second whether certain time slots for grocery delivery can still fit into the schedule. This model was created within the CILOLAB project and is seen as a significant step in optimising grocery delivery. [2] Although other parties have shown interest, the model has not yet been implemented at Albert Heijn. Van der Hagen states: ‘There are still steps to be taken before it is applied in daily practice, but the potential is certainly there.’ [1]
Impact on the Logistics Sector
The use of machine learning in grocery delivery not only affects customer satisfaction but also the efficiency of the logistics sector. By making faster and more accurate decisions, companies can serve more customers per trip, leading to a reduced ecological footprint and lower costs. Additionally, the model can help optimise routes and manage peak times, which is essential in a sector increasingly dependent on fast and reliable deliveries. [1][2]
Future Challenges
While the AI model is promising, it also presents challenges. One of the most important is ensuring due diligence when using self-learning systems. Lawyers such as Michelle Vrolijk from ITL Attorneys Netherlands point out that there will be more legal scrutiny of automation in the coming years. ‘Ensure that due diligence is adequately met now,’ advises Vrolijk. [3] Moreover, attention must be paid to the privacy and security of customer data used to train the model. [alert! ‘Privacy and security aspects are crucial, but details about implementation are limited.’]