A game theoretical model for profit maximization resource allocation in cloud environment with budget and deadline constraints
A game theoretical model for profit maximization resource allocation in cloud environment with budget and deadline constraints
复制标题
DOI：
10.1007/s11227-016-1782-z
复制
发表时间：
2016-06
期刊：
影响因子：
--
通讯作者：
A. Nezarat;G. Dastghaibyfard
中科院分区：
文献类型：
--
作者：
A. Nezarat;G. Dastghaibyfard
云环境中资源分配的一个重要挑战是基于云经济参数的定价问题和云资源申请者的选择。考虑到云环境中的资源分配是一个基于经济供需的问题，基于经济学的方法可以在更短的时间内获得更好的解决方案。本文利用贝叶斯方法，在每个用户估计其他竞争对手的行动，在拍卖的下一步，一个博弈模型的赢家确定。在不完全信息环境下，提出了一种基于组合拍卖的非合作博弈机制，以达到纳什均衡点并选择获胜者。使用所提出的方法，云提供商的利润提高了17%，销售资源增加了12%。所提出的投标目标函数在所有情况下都收敛于解，并且是稳定的。最后证明了所提出的模型具有获得当地最优报价的可能性。
One significant challenge for the resource allocation in cloud environments is the pricing issue and selection of applicants of cloud resources on the basis of cloud economic parameters. Taking into account the fact that the resource allocation in cloud environments is an economic supply- and demand-based problem, economics-based methods result in better solutions in a shorter period of time. In this paper, using Bayesian method, where each user estimates other rivals’ actions in the next step of the auction, a game model for winner determination is proposed. a non-cooperative game theory mechanism based on combinatorial auction in an environment with incomplete information has been proposed to reach Nash equilibrium point and select the winners. Using the proposed method, an improvement of 17 % profit was obtained for the cloud provider and a 12 % boost was seen in the sold resources. The objective function suggested for bidding converged to the solution in all cases and was stable. In the following, it was proved that the proposed model has the possibility of attaining the best local bid.