A closer look at the 6 experiments
The manufacturer of tailor-made windows and doors worked with a team from delaware to optimize processes for loading orders onto their delivery lorries. Their aim was twofold: simplifying the work of planning deliveries, and helping operators when loading lorries. To this end, an optimization algorithm was developed to improve the positioning of goods on the racks and the number of racks on the lorries. Mission accomplished!
As a designer of interiors and construction solutions, UNILIN manufactures a wide range of products in a variety of shapes, sizes and weights. Here too, delaware's mission was to optimize the way their products were loaded onto delivery lorries and to avoid any problems of excess weight during the loading process. However, some limitations could not yet be overcome by the quantum computing solution that was used. The experiment will therefore continue in 2024.
Nanoelectronics and digital technology R&D hub imec handles a vast amount of information. The challenge here was to improve the performance of their enterprise search, using generative AI (genAI), to help their 5,000 employees find information quickly. The initial results are promising, and the next step will be fine-tuning the process to make the results even more relevant. Another project to keep an eye on.
Active in the agri-food sector, COSUCRA extracts fiber from chicory root. This fiber, inulin, is then used in food processing. COSUCRA’s R&D department is aiming to develop new types of roots in order to maximizes inulin yields. A painstaking task, as proven by the 6,000 images and other documents accumulated over the years. The aim of the experiment is to use a computer vision algorithm and AI to extract the characteristics of each root, in order to facilitate the work of breeders. The results are promising, and the experiments will continue in 2024.
As an air-conditioning expert, TRANE aimed to be able to anticipate maintenance to be carried out on its customers' machines and thus reducing the number of visits. This experiment was carried out in close collaboration between students from Louvain School of Management (UCLouvain) for the business side (cost/benefit analysis), and delaware for the technical side. Their aim was to gather raw data from the machines and then use a predictive algorithm to determine the best time to intervene. As a result, TRANE will be alerted each time a machine needs servicing.
/DelawareDEL20_063-(1).webp?mode=autocrop&w=560&h=560&attachmenthistoryguid=6b7ed0cc-7a7d-47d3-b9ee-ba57aed3602c&v=&c=a345f87f8811b8f7f8609a89bc161aeb7ec84efd3c973ed84738bea36d6b2345)
.svg?lang=en-US&ext=.svg)
