The proper composition of activities is important for the design ofefficient and understandable workflow processes. Due to a lack ofavailable support, activity composition is currently a time-consuming,manual task that requires expertise and case knowledge. This thesispresents an automated approach that addresses this issue. The ap-proach utilizes the data-flow underlying a workflow process to de-termine the importance and semantic relatedness of the various, ele-mentary data-processing steps. Based on these aspects, fundamentalguidelines are proposed to drive and objectify the task of activitycomposition in the context of workflow design. A quantitative evalu-ation indicates that the use of these guidelines leads to activities thatclosely resemble those designed by experienced modellers. The guide-lines have been implemented in software that has been made freelyavailable. This enables any modeller to generate proper activities fora given workflow, irrespective of their familiarity with the businessprocess or the concepts of activity design.
Dieser Datensatz wurde nicht während einer Tätigkeit an der Universität Mannheim veröffentlicht, dies ist eine Externe Publikation.
Search Authors in
