中文摘要：

任务调度是云计算研究中的NP难优化问题,负载均衡问题是任务调度的热点研究内容之一。针对云计算环境中任务分配不够合理、用户服务质量低的问题,提出一种模拟蜜蜂采蜜机理的负载均衡策略。该策略模拟蜂群觅食行为,建立负载均衡模型,被迁移的任务作为侦察蜂更新虚拟机的负载信息,并采用贝叶斯分类算法对虚拟机负载状态进行分类,将任务从重负载虚拟机迁移至轻负载虚拟机,同时满足目标虚拟机中高优先级任务的数量最少,避免了大量任务被调度到同一性能较优的虚拟机上,能够有效减少任务等待时间。实验结果表明,基于蜜蜂采蜜机理的负载均衡策略与传统算法相比,减少了任务响应时间、完工时间和迁移次数,同时更好地满足了用户服务质量需求。

英文摘要：

Task scheduling is an NP-hard optimization problem for cloud computing research,and load balancing problem is one of the hot researches of task scheduling. For the problem that task allocation is not reasonable enough and user＇s quality of service is low in cloud computing environment,this paper proposed a load balancing strategy by simulating bees hunting honey mechanism. The strategy simulated the foraging behavior of honey bees,then established load balancing model. The migrated task as the scouter updated the information of VMs. The paper classified the VMs by Naive Bayes algorithm. It migrated the tasks in heavy loaded VMs to the light loaded VM whose number of tasks with higher priority in the queue was minimal. It avoided numerous tasks being scheduled to the same VM with good performance. This can effectively reduce the waiting time of the tasks. The experimental results illustrate that load balancing strategy based on bees hunting honey mechanism can reduce the response time,makespan of the tasks and migration times. At the same time it can meet the users＇ need of quality of service better.