An efficient approach for supporting dynamic evolutionary change of adaptive workflow

Abstract

Transferring multiple active instances from an original workflow model to a new one remains a big challenge for dynamic evolutionary change. Few approaches have been proposed to propagate model change to running instances. Existing ones adopt so strict criteria that too many transferable circumstances are left unmigrated. This paper proposes an efficient approach to smoothly migrate running instances to modified model. The approach is based on Dynamic Change Region (DCR) and Reduced Trace Set (RTS), which are introduced to identify transferable instances according to a newly proposed correctness criterion. A migration algorithm is provided to perform marking adaptation after transfer, and several examples are given for illustration. © 2008 Springer-Verlag Berlin Heidelberg.