Data challenge

Data-driven Logistics and Supply Chain Competition INCOM 2024

Description

This is a logistics and supply-chain data competition organised within INCOM 2024 (https://www.incom2024.org/), open for all participants registered for the conference. Participants can register as a team or as individuals through this link: https://forms.gle/67ixpD3Fy8uEH7H57 The exact problem and dataset will be made available to the participants a day before the conference commences.

The competition comprises two segments, one for a logistics problem and the other for a supply chain problem. Participants can choose to participate in both or either one of the segments. There will be a winner in each segment. These are further detailed below:

 

  • Segment 1: Logistics competition

    • Problem: This segment includes solving a dynamic vehicle routing problem. Logistics industries often face erratic customer demand, making it difficult for service providers to forecast and cater to the demands efficiently. In this segment, participants will be presented with a goal to optimise routes of two trucks which leave the depot daily with given demands from 15 customers at the start of the day. However, the customer demands are updated every 4 hours, and affect the original routes that the trucks are allocated at the beginning of the day. Customer demand is generated using data informed distributions, and therefore are not random.
    • In summary, the problem is to forecast the daily customer demand and efficiently update truck routes based on the observed demand.
    • Testing: Synthetic historical data shall be provided to the participants to develop their model and codes. An executable version shall be uploaded by the participants on the website for testing, using synthetic test data hidden from the participants for the duration of the competition.
    • Evaluation metric: An ideal routing algorithm shall ensure the customer orders are delivered with minimal delay and be cost and carbon efficient by minimising the distance travelled. The exact metric shall be unveiled on the first day of the conference.

 

  • Segment 2: Supply Chain Competition

    • Problem: This is a delivery delay prediction problem, often encountered by industries relying on more than one supplier. Here the participants assume the role of procurement professionals in a manufacturing industry with three tier one suppliers, which may or may not in turn have more suppliers. The supply network itself is static, however the configuration of the entire supply network will be unknown to the participants. Delivery delays, known to the participants, will be modelled by patterns drawn from real data.
    • Testing: Synthetic historical data consisting of delivery delays shall be provided to the participants to develop their delay prediction models. An executable version shall be uploaded by the participants on the website for testing, using synthetic test data hidden from the participants for the duration of the competition.
    • Evaluation metric: An ideal delay prediction algorithm shall be able to accurately predict and incorporate delays in the incoming materials

     

    In addition to the winners in each segment described above, a special prize will also be reserved for the most innovative algorithmic approach.

    Winners will be announced during the conference closing session.

    Competition data and details will be uploaded to the INCOM 2024 Website on the opening of the conference.

     

    Organisers:

    Prof. Alexandra Brintrup, University of Cambridge

    Dr. Liming Xu, University of Cambridge

    Dr. George Baryannis, University of Huddersfield

    Prof. Dmitry Ivanov, Berlin School of Economics and Law

    Please contact Dr Liming Xu (email: lx249@cam.ac.uk) for any questions.