QoS scheduling of fuzzy strategy grid workflow based on the bio-network

摘要

Grid workflow may construct new value-added services by composing existing elementary services with sophisticated workflow logic. Grid workflow schedule requires sophisticated expertise of domain in order to automate the process of managing workflows and its components. Quality of service (QoS) becomes essential and poses great challenges to grid workflow. QoS-aware selection and execution of grid workflow can fulfil customer's expectations. A fuzzy strategy modelling approach is proposed to design the grid workflow schedule, which is inspired by the characteristics of emergence and self-evolution in biological immune system. The grid workflow matching relation is achieved by bio-entities (mobile agents) negotiation based on projection-join closure to capture such message situations. We propose a fuzzy preferences evolutionary algorithm to resolve the QoS-aware selection optimisation, which enables the construction of QoS-aware workflow at both abstract and concrete service levels. The simulation results show that the approach can significantly improve performance. It adapts well to the changes of environments.