A Context-Aware Framework for Business Processes Evolution

Abstract

Run-time adaptability is a key feature of dynamic business environments, where the processes need to be constantly refined and restructured to deal with exceptional situations and changing requirements. The execution of such a system results in a set of adapted process variants instantiated on the same process model but dynamically restructured to handle specific contexts. Process evolution exploits the information on process variants to identify the best performing recurring adaptations and adopt them as general solutions in the process model. However, process variants are strictly related to specific execution contexts and cannot be adopted as general solutions. We propose a framework supporting context-aware evolution of business processes based on process instance execution and adaptation history. Instead of looking for recurring adaptations, we propose to look for recurring adaptation needs (i.e., process instances with the same context constraint violation and system configuration). Based on the analysis of adapted instances, we automatically construct and rank corrective evolution variants which can handle the problematic context. At the same time, we try to identify preventive evolution variants by constructing process variants which can prevent the adaptation need. We demonstrate the benefits of our approach using a car logistics scenario.