EXPlainable and Learning production & logistics by Artificial INtelligence (EXPLAIN)

Important results from the project

EXPLAIN aimed to boost profitability, sustainability, and competitiveness in Swedish manufacturing by integrated virtual production modeling with AI for decision support, focusing on energy and resource efficiency. EXPLAIN tackled this goal with three innovations: automatic virtual model generation, explainable AI algorithms, and knowledge management. The consortium comprised diverse partners, fostering sector-wide applicability. The project demonstrated human-machine co-learning, enhancing multi-objective optimization.

Expected long term effects

EXPLAIN aimed to fuse virtual production modeling with machine learning for decision support in production systems, focusing on multi-criteria decisions including energy and resource efficiency. Notably, successful cases with Scania, Seco-Tools, and AstraZeneca demonstrated the efficacy of this approach. These included optimizing productivity and energy consumption in Scania´s synchronizer production line and prototyping a lead-time forecasting tool for AstraZeneca´s quality control labs. Many of these outcomes are published in 12 papers.

Approach and implementation

In terms of AI for real-time decision making and knowledge management, EXPLAIN successfully explored deep reinforcement learning for multi-objective problems and utilized AI platforms for knowledge graph construction. These achievements are expected to pave the way for future AI tool applications in factory settings, with several outcomes published in national/international conferences, particularly regarding Scania´s cases. The current results encourage future research and implementation that utilize MOO/AI for productivity and energy efficiency.

External links

This web site is hosted by RISE as the representative web of the project.