HUPI – Optimisation of Production Planning to Maximise the Number of Orders Shipped Within the Deadlines Set by the Customer
HUPI
Sector: Industry
Business Case
Optimised production planning, which provides operators with a continuous view (before each shift) of the production orders to be processed at each stage of the process (manufacturing, assembly, processing, shipping) in the coming days, as well as the interdependence between these stages. In this way, resources can be organised more efficiently and dynamically.
Objectives
Increase on-time delivery and higher quality compliance. Improve productivity: fewer adjustments and fewer incidents on the production line. Improve operational/logistics costs: transport combination → better anticipation of needs. Improve quality of work life through more proactive and less chaotic production management.
Use case
Using the daily updated list of orders and expected demand, along with production capacity, logistic deadlines, and real-time stock and machine/equipment setup status, the system employs an operational search algorithm (pressure-optimised with uncertainty consideration). Its goal is to find the optimal plan for the upcoming days that ensures on-time delivery to customers without requiring capacity expansion.
Infrastructure
Cloud
Technology
Machine learning and deep learning
Data
Customer/production orders (ERP and HUPI), stock level (ERP), logistic flows/movements (logistic software), production capacity and production processes (HUPI).
Resources
Human resources: multidisciplinary team: 1 technical director, 2 mathematician & data scientist, 1 full stack developer, 1 frontend developer and expert digital UX team. Hupi Elastic Cloud Platform
Difficulties and learning
The major difficulty was user acceptance/implementation. We have had to adapt the model to their current way of working, adding more restrictions.
KPIs (business impact and metrics of the model)
Number of “kits on time” per month. Reduction of production line incidents. Economic value of savings generated in person-profile-hours/centre.
Funding
Private. The customer paid for the development of this application.
Collaborators, Partners
Own resources. It was a product made by our data team.
