Imagine a digital assistant that doesn't just give you a schedule, but can explain exactly why it chose that path and prove it won't fail. It combines the 'gut feeling' of modern AI with the strict rule-following of traditional logic. This ensures that complex plans for factories or flights are both efficient and safe.
Current AI planning tools are often 'black boxes' that lack the transparency and safety required for critical industrial use. This prevents companies from fully adopting automated scheduling due to risks of unpredictable or unexplainable failures.
A set of open-source software tools and test environments. These include hybrid AI methods for planning and scheduling, hosted as use case demonstrators on GitHub.
If you are an airline operations center dealing with flight diversions — this project developed a hybrid AI system that ensures flight schedules are robust and explainable. This helps operators trust automated decisions during disruptions.
If you are a smart factory operator dealing with complex production sequencing — this project developed preference-learning methods that reduce the need for expert manual input by up to 72%. This speeds up the setup of production schedules.
If you are a grid management company dealing with demand uncertainty — this project developed a robust optimization approach that is two orders of magnitude faster than previous methods. This allows for quicker reactions to energy fluctuations.
Based on available project data, no pricing information is provided as the project focuses on developing open-source tools and test environments.
The project advanced technology from TRL2 to TRL4, meaning it has been demonstrated in laboratory settings across five use cases, but is not yet at full industrial scale.
The project aims to release open-source software tools and test environments to enable wider development and assessment.
It integrates symbolic methods with data-driven AI to provide verification and explainability, ensuring that decisions are safe and transparent.
Based on available project data, specific integration timelines are not provided, though the project developed tools to support decision-support systems.
The consortium is well-balanced for technology transfer, consisting of 9 partners across 7 countries. With a 44% industry ratio (4 companies, including 1 SME), the project has a strong link to practical application, while 5 universities provide the deep research needed for the hybrid AI components.
Contact LA COMMUNAUTE D UNIVERSITES ET ETABLISSEMENTS DE TOULOUSE
Contact us to explore the open-source TUPLES tools for your scheduling needs.
