Architecture for ontology-supported multi-context reasoning systems

Modern smart systems such as those needed for Industry 4.0 integrate data from various sources and increasingly require that data be contextualized with domain knowledge. The integration and contextualization of data allows for the advanced reasoning needed to generate knowledge grounded in the data under consideration. In this paper, we propose an architecture for an ontology-supported multi-context reasoning system which inherently supports a number of desired system qualities including data transparency, system interactivity, and graceful aging. The architecture is inspired by the Presentation–Abstraction–Control architecture style, which is an interaction-based architecture. Our architecture uses a two level hierarchy with three agents and can incorporate and utilize multiple contexts. It is flexible, supporting an interface between data and users, highly interactive, and easily maintained. The evolution of data is isolated to a single component of the system and therefore does not cascade to several others. A domain of application can be easily determined by the use of archetypes and domain-specification components. Our architecture is demonstrated using a case study involving data from the city of San Francisco.