Research On Pricing And Scheduling Considering Multi-party Benefits In The Edge Service Market

The development of cloud computing and edge computing has given birth to the computing model of "edge service market".With the increase of service requests,the situation of large-scale concurrent requests is inevitable,and the failure of requests is gradually increasing.In severe cases,the system may crash.Therefore,the users' request strategies need to be optimized.How to optimize users' request strategies while improving user utility has become an important research problem.And the providers also optimize their own utilities.The goals of both parties are contradictory.Price is an adjustment lever in the edge service market.Specifically,if the service price is too high,the provider's utility is great,but the user's utility is low because of the high cost.If the service price is low,the user's utility is great,but the provider's utility is low due to the low profit.This thesis mainly studies the problem of multi-party interest balance from two aspects:the optimization of request strategy and pricing mechanism.Aiming at the problem of large-scale concurrent requests of multiple users under edge computing,this thesis proposes a request strategy optimization mechanism based on non-cooperative games.The traditional task scheduling problem is transformed into the problem of how to maximize the profit of the provider and the payoff of the user.A multi-party collaboration benefit game model is constructed to optimize the user's request strategy and solve the problem of large-scale request concurrency.The distributed task outsourcing algorithm(DTOA)is designed to obtain the optimal strategies for request and service price.The experimental results show that DTOA can increase the profit of the provider and the payoff of the user,while also reducing the peak load of the system.With the popularity of the Internet of Things and the deployment of a large number of sensors,data has exploded.Big data is regarded as a kind of"commodity",which is used to extract useful knowledge and information and provide it to consumers as a data service.It forms a data market.This thesis aims to discuss three issues:how to price raw data,how to price data service,and how to optimize the interests of market participants.It focus on these three issues.First,a three-tier model of big data trading market is established,including a data provider,a service provider,and users.A method to quantify the value of raw data using two dimensions is designed.Secondly,the threeparty transaction behavior is simulated as a Stackelberg game,and a multiparty linkage big data pricing mechanism is formulated.Finally,the experimental results show that the three parties' interests can be balanced through the proposed pricing mechanism.