Profit-driven Service Request Scheduling For Heterogeneous Users In Cloud Computing

As people's demand for computing resources grows,cloud computing becomes more and more popular as a pay-as-you-go model that can efficiently provide computing resources and services,and cloud computing providers are also emerging.The typical goal of cloud computing providers is to maximize their profits.The main goal of users is to obtain their own satisfactory service prices and quality of service,and each user has their own preferences for service prices and quality of service.As a provider of services,cloud computing providers should understand the personalized preferences of service prices and quality for users with different personalities,and assign and schedule service requests based on their personalized preferences to maximize profits under user satisfaction constraints.This paper considers the user's personalized preferences,designs a novel real-time service request scheduling algorithm,and optimizes the profit of the cloud computing provider under the constraint of user satisfaction.Specifically,1.This paper proposes the concept of user personality and achieves accurate modeling of the satisfaction of different personality users.This paper establishes a statistical model to estimate the user's satisfaction with the given service price and service quality.The statistical model can predict the personalized demand of service price and service quality for users with different personalities,so that users can get satisfactory service.2.This paper designs a cloud provider profit optimization scheme based on Integer Linear Programming(ILP).In this paper,based on ILP,the allocation and scheduling of personalized user service requests are modeled and solved to obtain the optimal profit scheme of cloud providers under the constraint of user satisfaction,and the shortcomings of ILP-based methods are proposed.3.This paper proposes a heuristic profit optimization algorithm based on value evaluation and cross entropy.In order to overcome the shortcomings of the ILP-based method,this paper proposes a heuristic profit optimization algorithm for real-time allocation and scheduling of personalized user service requests.The algorithm can obtain the approximate optimal allocation and execution order of the user service request,so as to realize the profit optimization of the cloud provider under the constraint of user satisfaction.Experimental results show that the user satisfaction model proposed in this paper can accurately predict the user satisfaction score.Compared with the other three existing service request scheduling schemes,the proposed ILP-based profit optimization scheme can achieve an average profit increase of up to 97.06%,42.45% and 23.58%,and the proposed heuristic algorithm can achieve an average profit increase of up to 84.76%,31.22% and 10.35%.