Research On Resource Allocation Technology Based On Multi-Period Network Service Request Policy Learning

In the 5G network era,business requirements and application scenarios are gradually diversifying.It is unrealistic for operators to create different private networks for different services,and the cost of construction,operation and maintenance will be very high.There are huge challenges in distribution.Therefore,how to reasonably allocate network resources and improve the income of network operators will have important research significance.In response to this problem,this paper proposes a network resource allocation technology based on the analysis and learning of the resource requirements of network service providers.Through reasonable pricing of network resources,the income of network operators is optimized,and the price is leveraged to achieve reasonable allocation of network resources.the goal of.In general,because the network service provider has a request strategy,its resource requirements are uncertain.The network operator needs to perform statistical analysis and learning on the resource requirements of the network service provider,and divide the cycle of providing network resources into two periods,and balance the supply and demand of network resources through the pricing strategy of network resources.Firstly,in response to the problem of resource overload caused by the concentration of demand for network resources from network service providers,this paper proposes a network resource allocation technique based on demand analysis response pricing.Specifically,the network resource supply and demand problem between network operators and network service providers is modelled as a two-stage dynamic game,and the resource demand of service providers in each stage of the network segment is analysed to determine the utility function of network service providers;the resource demand in the first stage is combined with the prediction analysis of the resource demand in the second stage to determine the utility function of network operators;then,the inverse induction method is used to The Nash equilibrium is solved for the game model to determine the pricing strategy and expected equilibrium revenue of the network operator in each stage of the equilibrium.Finally,the simulation results show that the strategic behaviour of network service providers has a negative impact on the equilibrium revenue of network operators,and the higher the level of their strategies,the greater the impact on the equilibrium revenue of network operators,but by incentivising network service providers to delay the demand for resources through response pricing,the supply and demand of network resources can be effectively balanced and resource utilisation improved.Then,to address the negative impact of the network service provider’s request strategy on the network operator’s equilibrium revenue,this paper proposes a demand analysis commitment pricing-based network resource allocation technique.First,the network operator’s first-stage commitment pricing strategy is analysed by the demand strategy of the network service provider to determine the utility function of the network operator;then,the shortcomings of this commitment pricing strategy are analysed and a pricing scheme combining response pricing and commitment pricing is proposed.Finally,the simulation results show that the commitment pricing strategy can effectively control the excessive request delays of network service providers and improve the overall revenue of network operators.Finally,through numerical simulation and analysis,the expected equilibrium revenue of network operators in the above two network resource allocation models are compared.The results show that when facing network service providers with higher policy levels,the response pricing strategy can incentivise network service providers to delay demand,but the equilibrium revenue of network operators is lower;while the commitment pricing strategy can control the request strategy of network service providers and improve the equilibrium revenue of network operators.