可分割云资源下多租户效用最优化策略
发布时间：2019-01-20 12:44
【摘要】：为有效解决多租户对可分割云资源分配的竞争问题,提出一种基于效用函数的最优化分配策略。通过定义竞争租户对资源分配的评估函数和代价函数,建立效用最优化模型。讨论不同形式的代价函数对最优化分配结果的影响,验证在该效用模型下资源分配存在纳什均衡解,仿真对比不同形式代价函数下资源分配的性能表现。实验结果表明,在求解租户社会福利最大化过程中,线性代价函数和凸性代价函数均可以使分配过程收敛到唯一解上,但线性代价函数可以为单个租户带来最优的系统效率,实现多租户环境下整体社会福利的最大化。
[Abstract]:In order to solve the problem of multi-tenant competing for partitioned cloud resource allocation, an optimal allocation strategy based on utility function is proposed. The utility optimization model is established by defining the evaluation function and cost function of resource allocation for competing tenants. The effects of different cost functions on the optimal allocation results are discussed. It is verified that there is a Nash equilibrium solution for resource allocation under the utility model. The performance of resource allocation under different form cost functions is compared by simulation. The experimental results show that both linear cost function and convexity cost function can converge to the unique solution in the process of maximizing the social welfare of tenants, but the linear cost function can bring the optimal system efficiency for a single tenant. To maximize the overall social welfare in a multi-tenant environment.
【作者单位】： 黄河科技学院现代教育技术中心;河南师范大学计算机与信息技术学院;
【基金】：河南省科技厅研究基金项目(20122015)
【分类号】：O224
本文编号：2412052
[Abstract]:In order to solve the problem of multi-tenant competing for partitioned cloud resource allocation, an optimal allocation strategy based on utility function is proposed. The utility optimization model is established by defining the evaluation function and cost function of resource allocation for competing tenants. The effects of different cost functions on the optimal allocation results are discussed. It is verified that there is a Nash equilibrium solution for resource allocation under the utility model. The performance of resource allocation under different form cost functions is compared by simulation. The experimental results show that both linear cost function and convexity cost function can converge to the unique solution in the process of maximizing the social welfare of tenants, but the linear cost function can bring the optimal system efficiency for a single tenant. To maximize the overall social welfare in a multi-tenant environment.
【作者单位】： 黄河科技学院现代教育技术中心;河南师范大学计算机与信息技术学院;
【基金】：河南省科技厅研究基金项目(20122015)
【分类号】：O224
【相似文献】
相关期刊论文 前3条
1 刘晓辉;党亚民;王潜心;杨磊;;双基线姿态确定三种算法的比较分析[J];大地测量与地球动力学;2013年01期
2 徐奕奕;唐培和;刘浩;;基于消费代价权衡的网格动态资源分配算法[J];科技通报;2012年02期
3 ;[J];;年期
相关会议论文 前1条
1 朱宏;白文乐;牛长流;;PAM调制方式下定时估计算法中代价函数的性能分析[A];2008通信理论与技术新发展——第十三届全国青年通信学术会议论文集（下）[C];2008年
相关硕士学位论文 前3条
1 岳鹏;基于矩阵嵌入的自适应VoIP隐写方法研究及实现[D];南京理工大学;2017年
2 张鹏辉;基于代价函数的路由策略设计原则分析及算法改进[D];中南大学;2013年
3 王明;最小化嵌入失真图像隐写的代价函数设计[D];深圳大学;2015年
，本文编号：2412052
本文链接：https://www.wllwen.com/kejilunwen/yysx/2412052.html
