Real-Time Context Monitoring and Analysis for Detecting Process Adaptation Needs

The increasingly dynamic and changing environment in which business processes are evolving requires companies to adapt them frequently. Thus, we deal in this paper with an adaptation engine ensuring the adaptation need detection of running processes, and more precisely of its first two components: the Monitor (M) and Analyze (A) components. This adaptation engine implements the MAPE-K loop and uses a contextual approach for the detection of the adaptation needs. It is based on a model that enables a complete representation of the process operating environment (OE) using the context notion. More precisely, this paper presents the architecture of the MAPE-K-based adaptation engine and introduces the BPMN4Context meta-model, which supports the modeling of the operating environment using the context along with the model of all processes and their use conditions. In addition, the paper introduces the recommended context-based approach that advocates (i) a filtering activity to select only significant context changes in the monitored data, considered as low-level context parameters, (ii) a reasoning activity to deduce high-level context parameters from filtered low-level ones, enhancing the current situation of running processes and (iii) the examining of the current situation before its analysis in order to resolve problems related to the used units and synonym values. Our approach is implemented as a system that instantiates the M and A components of MAPE-K loop. The feasibility and applicability of this approach is demonstrated by a case study from the crisis domain, a set of criteria and two performance tests.