Migrating workflow process fragmentation based on cost constraint and balanced granularity

In migrating workflow system, process fragmentation is to partition a process into a set of sub-processes which can be allocated to the migrating instances for the process execution. Due to the distributed and constrained workplace resource, process fragmentation needs take the execution environment into consideration. In addition, the balanced granularity of the fragmentation also need to be taken into account so as to cater for the applications in which each migrating instance is generated with a balanced executing ability. This paper proposes a two-phase approach to the process partitioning. First, we abstract the process resource allocation as a multi-objective combinatorial optimization problem and present a particle swarm optimization algorithm for the solution of the problem. And then, based on the mapping relationship between the tasks and the available resources, we propose a task clustering algorithm to generate the set of sub-processes with balanced granularity. Simulation experiment results show the effectiveness and feasibility of the fragmentation approach.