Predictive Analytics – Opportunities for Factory Automation
Prof. Dr. Wolfgang Lehner
Director Institute of System Architecture
Dresden University of Technology (TUD)
Detailed predictions of production process KPIs are crucial for making decisions regarding production quotas, maintenance schedules or capacity planning. At present, most planning processes rely heavily on complex simulations that require manual tuning and are made to order.
However, the substantial amount of data that is already collected from tools, sensors and other systems in todays automated production environments offers promising opportunities for data-driven approaches to support planning and decision making.
In this presentation, we introduce some of our approaches for large-scale predictive analytics: a robust and adaptable algorithm for large-scale forecasting, a system for on-demand KPI forecasting, and a modeling approach for what-if scenarios allowing the assessment of possible future developments. We outline each of these approaches and point out the insights they can offer regarding planning and decision making.
- Full Professor for Database Systems at TU Dresden, Germany
- Visiting Scientist at
- SAP Labs, Wallorf, Palo Alto, …
- Microsoft Research, Redmond (WA)
- IBM Almaden, San Jose (CA)
- University of Waterloo
- Chairman of the DFG Computer Science Reviewing Board
- Member of the Academy of Europe
- Member of the VLDB Endowment Board of Trustees
- Member of the Board of Trustees of the Saxonian State Library (SLUB)