Paschke, A.A.Paschke2022-03-142022-03-142016https://publica.fraunhofer.de/handle/publica/408080In this paper we report on a predictive maintenance use cases using Provalet rule agents for implementing expressive rule-based streaming analytics and decision logic on top of online machine learning prediction models, which are dynamically applied to the streaming data coming from on-board asset monitoring sensors. Provalets are component-based mobile agents for rule-based inference analytics, which can be dynamically deployed as microservices into container environments via simple REST calls.en004conference paper