Alex de Mulder // Giovanni Pazienza
(keywords: knowledge discovery, scheduling, multi-agent system)
MIDAS is a multi-agent solution to handle unexpected (and potentially disruptive) events during the scheduling of some operations (currently, it has been applied to manufacturing and health management domains). Basically, it elaborates statistics and performs data mining on some data gathered by sensor nets or by other means (UIs, historical data, etc.). The information inferred by MIDAS can be used to refine the results of the scheduler and to understand possible causalities of unexpected events.
MIDAS is based on Eve (for the agent part) and vis.js (for the visualization part). Currently, it displays the results in form of bar charts (above) or timelines (below).
Estimated TRL/Maturity: TRL 5 (i.e., technology validated in relevant environment), medium maturity (tests and documentation are still needed)
Past projects: ARUM
Unique selling points: open source, high flexibility, possibility to be fed with real-time data (connection with the internet of things), knowledge discovery
Competitors: Unknown at the moment
Key references: ARUM documentation
Demos available: ARUM use case demo's, NHS use case demo
Possible developments: further documentation, possible integration with ASIMOV, possibility to extend to other domains
Ideas for application: energy and mobility projects in Smart Cities (this can be used as a tool for scheduling and rescheduling of events), factory of the future calls (applications in manufactoring), or anything that has processes and data/measurement in it
Alex de Mulder