Creation of data-driven decision platform to manage the flow of reusable crates
Overview
| Company Name / Department | Pool Service |
| Contact Person | Björn Joosten |
| Location | Eindhoven/Deventer |
| Optional remote work | 1x per week Eindhoven. Deventer to be agreed. |
| Travel expenses (own account or reimbursed by the company) | Reimbursed |
| Housing arranged by company | No |
| Housing expenses (how much per month, own account or subsidized by the company) | Not applicable |
| Internship compensation | € 750 per month based on full-time |
| Study program | Industrial Engineering department |
| Start date | September 2026 |
Company Description
At Poolservice, we are responsible for managing a return logistics concept for a large pool of crates that are used at nearly all Dutch retailers (Jumbo, AH, Plus, etc.) in which products are shipped from suppliers towards the retailers.
Project Description
To manage this pool of crates, Pool Service is developing an advanced data-driven decision-making platform that will leverage retailer and supplier specific characteristics to forecast the different streams of crates throughout its network.
The goal of this project is to develop an advanced data-driven decision-making platform that will leverage retailer and supplier specific characteristics to forecast the different streams of crates throughout its network, and to optimize the allocation of returned crates from retailers towards suppliers.
This requires:
- Several forecasting modules on different parts of the crate network including the demand from retailer at suppliers, the returned crates from retailers, and the demanded crates from suppliers at poolservice.
- An allocation rule on how to ship the returned crates towards suppliers.
We specifically looking for a solution that provides daily decision-support, and encourage the development of a tool that can help our planners in their daily operations.
Goal of the Project
Develop a data-driven decision-support platform that forecasts daily crate flows — including retailer demand, crate returns, and supplier orders — across Pool Service’s network of 100+ nodes. These forecasts feed an allocation model that determines how returned crates should be routed back to suppliers each day.
Deliverables
Models and validation for forecasts, supported by Bachelor project that will run in parallel.
Allocation model and decision ruling/optimization.
Proof-of-concept prototype model.
Essential student knowledge
Interest in or exposure to forecasting methods
More information: escf@tue.nl
