Research On Intelligent Resource Management Algorithm For Delay Optimization In Network Slicing

In 5G era,facing with the explosive growth of communication services,the network should be able to simultaneously support a large number of diverse traffic scenarios from vertical industries,therefore,network slice technology emerges at the historic moment under this kind of background.The network slice technology could provide customized services for different traffic scenarios to meet their differentiated performance requirements by means of creating multiple logical virtual networks on one physical infrastructures.Considering the limitation of physical resources and the dynamics of service requests,unreasonable resource allocation would increase system delay while decrease users' experience.Therefore,it is extremely important to establish a novel and efficient intelligent resource management mechanism,which can dynamically adjust and optimize resource allocation strategy according to the dynamics of environment.In consequence,the target of this thesis is to optimize resource management strategy in network slicing,and the main research works and innovations are summarized as follows:To solve the problem of high system delay caused by unreasonable resource allocation because of randomness and unpredictability of service requests in 5G network slicing,a slice resource management scheme based on transfer Actor-Critic(A-C)learning is proposed.Firstly,a system delay minimization model based on VNF placement and joint allocation of computing resources,link resources and fronthaul bandwidth resources is built,then transformed into a discrete-time Markov Decision Process(MDP).Next,A-C learning algorithm is adopted in the MDP to dynamically adjust Service Function Chain(SFC)resource allocation strategy by interacting with environment,so as to optimize the system delay.Furthermore,in order to realize and accelerate the convergence of the A-C algorithm in similar target tasks(such as the arrival rate of service requests is generally higher),the transfer A-C algorithm is adopted to utilize the SFC resource allocation knowledge learned from source tasks to quickly find the allocation strategy in target tasks.Simulation results show that the proposed scheme can reduce and stabilize the queuing length of SFC packets,optimize the system delay,and improve resource utilization.Aiming at the slice resource management optimization problem caused by mobility of user equipment(UE)and dynamics of packet arrival in the radio access network slice,and considering the fact that global network information is hard to obtain but necessary in the process of optimization,a resource management scheme based on Asynchronous Advantage Actor Critic(A3C)learning is proposed.First of all,a resource management mechanism based on blockchain technology is established,which can credibly share and update the global network information,also to supervise and record SFC resource allocation process.Then,a delay minimization model based on joint allocation of radio resources,computing resources and bandwidth resources is built under the circumstance of UE moving and time-varying packet arrival,and further transformed into a MDP problem.At last,A3 C learning method is adopted to obtain the resource allocation optimization strategy in this MDP.Simulation results show that the proposed scheme can utilize resources more efficiently to optimize the system delay while guarantee the requirement of each UE.