Activity prediction in process management using the WoMan framework

Abstract

In addition to the classical exploitation of process models for checking process enactment conformance, a very relevant but almost neglected task concerns the prediction of which activities will be carried out next at a given moment during process execution. The outcomes of this task may allow to save time and money by taking suitable actions that facilitate the execution of those activities, may support more fundamental and critical tasks involved in automatic process management, and may provide indirect indications on the correctness and reliability of a process model. This paper proposes an enhanced declarative process model formalism and a strategy for activity prediction using the WoMan framework for workflow management. Experimental results on different domains show very interesting prediction performance.