Resource allocation mechanism based on neural network and Swarm search optimization in cloud environment

Resource allocation mechanism based on neural network and Swarm search optimization in cloud environment

Sun Jiajia Xingwei Gaocheng

In the cloud environment, a variety of idle resources can be pooled to form a pool of resources, and then use virtualization technology to the resource pool in the form of different resources to provide users with services, Therefore, a reasonable and effective mechanism is needed to allocate resources. According to the characteristics of the resources in the cloud environment, combining the economics and the intelligent method, a new intelligent resource allocation mechanism based on bidirectional combined auctions is proposed. In this mechanism, we propose a quality ofexperience based on the experience quality. The Prestige system of QoE), the prestige attenuation coefficient and the user's prestige are introduced to reduce the effect of malicious behavior in auction, and to provide QoE support for resource transaction. The bidding decision mechanism based on BP Neural network is proposed, which can not only reasonably determine the bid price, but also comprehensively consider various factors. And so that prices can be dynamically adapted to market changes. Finally, because the problem of combinatorial auction winning is NP-complete, the Group Search optimization algorithm is introduced to optimize the target of market surplus and overall prestige to obtain the resource allocation scheme. The simulation results show that the mechanism is feasible and effective.

Resource allocation mechanism based on neural network and Swarm search optimization in cloud environment