Context
Internal challenge: As part of a competition aimed at challenging the capabilities of AI to improve its processes and products, a Valeo team was accompanied by Datalchemy from conception to completion of a proof of concept on this subject.
Need
- Automate the generation of shift schedules and the assignment of production lines within an equipment manufacturing plant.
- Integrate all HR constraints (availability, skills), machine capacity, stock levels and delivery requests.
- Respond quickly and flexibly to changing workloads and unforeseen absences.
Completed work
- Business modeling: formalization of constraints and key concepts to structure the problem.
- Data enrichment: creation of synthetic data sets to supplement the files supplied by Valeo.
- Constraint solver: integration of Timefold to automatically produce optimized schedules according to various scenarios.
- Interactive prototype: development of a simple, functional demonstration interface, enabling results to be viewed, tested and adjusted in real time.
Results
- Operational base: a concrete proposal serving as the basis for a more comprehensive, industrializable tool.
- Transparency and anticipation: business modeling makes it easier to plan inventory and schedules in advance.
- Successful Proof of Concept: demonstration of the feasibility of a large-scale solution, paving the way for substantial gains in flexibility, responsiveness and performance.