摘要：Most existing task scheduling algorithms in cloud storage fail to aware users ' QoS preference. In addition, these algorithms result in low user satisfaction rate for they do not consider the characteristics of cloud storage. In order to address these problems, the "optimal order comparison method " is used to aware users ' QoS preference, and also helps experts use their professional knowledge to decide the weight of QoS classes. We redefined the fitness function of the particle swarm optimization (PSO) algorithm by using these weights and proposed the "PSO based hierarchical task scheduling with QoS preference awareness: PSO-HQoSPA" algorithm. By consider both user and expert experience, the method can aware users’ QoS preference and deal with multiple QoS requirements. The simulation results show our method achieved acceptable user satisfaction rate, and the same time maintains the efficiency as traditional PSO based method.
