关键词: 语义工作流, 并行化, 数据生成&mdash, 消耗依赖, 资源约束, 并行度

Abstract: The parallelism of business process models is an important indicator of its quality.The business process model with high parallelism has a better quality and a higher operating efficiency.Aiming at the requirement of improving semantic workflows qualities and the fact that existing methods of business processes model parallelism cannot deal with the data produce-consume dependency and resource constraints in semantic workflows,a parallelism refactoring algorithm of semantic workflows based on data produce-consume dependency was proposed.The task relations matrix of semantic workflow was obtained by using numbered nodes and the nearest common ancestor.The data dependency matrix was obtained by analyzing the data produce-consume dependency between task nodes.The update rules of task relations were proposed,with which the task relations matrix was updated according to the data dependency matrix to gain the task relations matrix based on data dependency.A refactoring algorithm based on the new task relations matrix was designed to generate the parallelized semantic workflow while giving consideration to resource constraints.A measure was proposed to assess the degree of parallelism for semantic workflows,while interactive software for parallelism refactoring of semantic workflows was implemented.The simulation results showed that the proposed algorithm could improve the parallelism and qualities of semantic workflows,and provide an effective support for improving the operating efficiency of business processes based on semantic workflows.

Key words: semantic workflow, parallelism, data produce-consume dependency, resource constraints, degree of parallelism