🕒 Article read time: 2 minutes
Generation Logistics case study: Christina Kouridi, AI Research Engineer, InstaDeep London
Christina graduated from university with a Masters degree and took her first step into logistics as a supply chain manager and consultant. But then she became interested in AI, and what machine learning could bring to the sector.
“I’ve always been drawn to hard problems,” she explains. “And bin packing is one of the hardest ‘unsolvable’ problems in computer science.”
Bin packing is an organisational problem, crucial to logistics, which focuses on how to fit goods of different sizes into the minimal possible numbers of containers.
“It’s theoretically an unsolvable problem,” continues Christina. “However, by applying re-enforceable learning we were able to get to solutions that humans would otherwise never be able to achieve.”
Christina finds these solutions through specialised software known as DeepPack.
“It’s AI-powered and helps to optimise loads so we can best pack a set of items in a container. The algorithm behind the software is unique. Firstly, because it’s the first to use machine learning to address the problem. And secondly, because it applies reinforcement learning on a really large scale.
“It’s the first of its kind.”
Published On: 07/03/2024 16:00:00
Comments Section
If you are a Logistics UK member login to add comments.
