Towards algorithmic generation of business processes: From business step dependencies to process algebra expressions

Abstract

Recently, a lot of work has been done on formalization of business process specification, in particular, using Petri nets and process algebra. However, these efforts usually do not explicitly address complex business process development, which necessitates the specification, coordination, and synchronization of a large number of business steps. It is imperative that these atomic tasks are associated correctly and monitored for countless dependencies. Moreover, as these business processes grow, they become critically reliant on a large number of split and merge points, which additionally increases modeling complexity. Therefore, one of the central challenges in complex business process modeling is the composition of dependent business steps. We address this challenge and introduce a formally correct method for automated composition of algebraic expressions in complex business process modeling based on acyclic directed graph reductions. We show that our method generates an equivalent algebraic expression from an appropriate acyclic directed graph if the graph is well-formed and series-parallel. Additionally, we encapsulate the reductions in an algorithm that transforms business step dependencies described by users into digraphs, recognizes structural conflicts, identifies Wheatstone bridges, and finally generates algebraic expressions. © 2009 Springer Berlin Heidelberg.