Research On WiFi Offloading Algorithms In Heterogeneous Networks

With the popularization of wireless devices such as smartphones and tablets,and the emergence of technologies such as virtual reality and cloud computing,people are becoming more dependent on mobile communication,which places increasing load on cellular networks.It brings many problems such as low communication rate,high session delay,and data discontinuity during peak hours.Wi Fi offloading uses Wi Fi networks to offload data on cellular networks.The load of cellular networks can be transferred to Wi Fi networks to solve the problem of the congestion of liscensed spectrum,which has caused widespread concern.This thesis studies the Wi Fi offloading technology in mobile networks.For the scenario where cellular base stations and Wi Fi access points coexist,a coalition game-based Wi Fi offloading algorithm,a Wi Fi offloading algorithm based on Q-learning and MADM(multiple attribute decision making)and a Wi Fi offloading algorithm based on decision tree and MADM are proposed.The main research contents and innovations of this article are as follows:(1)For the scenario where cellular network and Wi Fi network coexist,a Wi Fi offloading algorithm based on coalition game is proposed.Cellular networks and Wi Fi networks are considered as two types of coalitions.The cellular coalition optimizes resource block allocation,and the Wi Fi coalition optimizes access time allocation.Considering the user's communication rate,cost,and communication delay comprehensively,a utility function model that guarantees fairness is defined,and with the goal of maximizing the total utility of the system,a coalition transfer criterion that simultaneously improves the total utility of the system and the individual user's utility is established to control users to transfer between different coalitions.Simulation results show that the algorithm can converge in a limited number of times.Compared with traditional algorithms,this algorithm has improved the overall system utility.(2)For a mobile user scenario where cellular networks and Wi Fi networks coexist,combining reinforcement learning and multi-attribute decision-making,a Wi Fi offloading algorithm based on Qlearning and MADM is proposed.This algorithm not only considers the current network status,but also considers the user's accessing history,and uses Q-learning scheme to make the final offloading decision.The AHP algorithm is used to obtain the weights of the four attributes of user throughput,terminal power consumption,user cost,and communication delay.The TOPSIS algorithm is used to obtain the reward function in Q-learning.The user continuously updates his cumulative discount reward by combining instant rewards and experience rewards till convergence.Simulation results show that the user satisfaction of this algorithm is better than the traditional Wi Fi offloading algorithms.(3)Aiming at a heterogeneous network scenario where multiple cellular base stations and Wi Fi access points overlap,combining supervised learning with MADM,a Wi Fi offloading algorithm based on decision tree is proposed.The AHP algorithm and the GRA algorithm are used to calculate the recommendation priority of candidate networks and tag the samples.Using the generated training set and feature set to train the decision tree,users can make offloading decisions based on the classification results of the decision tree and choose the network with the highest recommendation priority to access,which can reduce the complexity of Wi Fi offloading.Simulation results show that,compared with the traditional Wi Fi offload algorithm,this algorithm guarantees the user's Qo S(quality of service)with lower complexity.