Developing a workflow design framework based on dataflow analysis

Abstract

One important step in workflow automation is to design a valid and efficient workflow model that accurately specifies the execution sequence of activities in a business process. The dataflow constraints, which define the data input and output of each activity in a business process, can be used to analyze the dependencies among different activities and determine a proper workflow model for a particular business process. The basic idea of applying dataflow analysis to workflow design has been suggested in the literature. In this paper, we present a framework for applying dataflow analysis to workflow design. Our contribution is threefold. First, we introduce several new instruments, such as activity dependency tree and condition-action table, for specifying the necessary workflow requirements. Furthermore, we develop various key algorithms for requirements analysis and workflow model generation. Third, a system architecture is provided as a roadmap for developing a dataflow-based workflow design tool. © 2008 IEEE.