This paper makes four main contributions towards managing large collections of process models. First, we show how flexible process variants can be configured by applying rules to a generic process template using a configuration algorithm. This leads to a separation of control flow and business policy. Second, we develop a new succinct representation for process trees as strings by performing a post-order traversal. This structure can facilitate process variant configuration and retrieval. Third, we develop techniques for querying a repository of process variants by means of bit vectors. In this way the management of variants even in a large repository is greatly enhanced. Finally, we describe a preliminary implementation of our approach with a case study for demonstration. Our focus is on capturing deeper process knowledge and achieving a holistic approach to robust process design, as well as ease of accommodating changes to business policy.
All Science Journal Classification (ASJC) codes
- Computer Science(all)