Research On Wireless Network Link Scheduling Problem Based On Random Geometry

As wireless communication technology gradually becoming mature,the application area of wireless networks is becoming wider and wider.What is different from wired networks is that several links in real wireless networks transmit concurrently and cause interference,so that the signals can't be coded normally,which leads to transmission latency and lowering transmission capacity.Link scheduling problem is a fundamental problem that needs to be solved in wireless network transmissions.Designing reasonable and effective link scheduling algorithms can reduce the interference produced by the links when transmitting concurrently,increase network throughput,lower transmission delay,and decrease transmission energy consumption.As a result,studying link scheduling problem,and getting correct and effective scheduling algorithms are of vital theoretical significance and application value.Recently,stochastic geometry emerges as a mathematical tool,it can analyze wireless networks with random topology and acquire its performance indexes.The power of stochastic geometry approach lies in that it is able to account for the randomness of nodes' location,meanwhile,it can accurately describe the physical layer based on SINR,also,it allows taking random fading into account on every link.Modeling wireless networks by stochastic geometry models and getting some performance indexes of the network is actually helpful for us to solve link scheduling problem.Based on analyzing existing results,the thesis mainly studies two typical problems of link scheduling in wireless networks,that is,single slot scheduling problem and shortest link scheduling problem.Besides,we use stochastic geometry approach to model a wireless network,and acquire important performance parameters of link scheduling problems.The detailed work and innovation points of this thesis mainly run as follows:Firstly,the thesis introduces stochastic geometry approach to link scheduling problems in wireless networks,and uses stochastic geometry models to obtain important performance parameters of wireless networks.We solve the successful transmission probability of a link in wireless ad hoc networks and cellular networks,transmission capacity and transmission delay of wireless ad hoc networks under Poisson Point Process.Also,we get their formulations and provide detailed theoretic proof to validate the correctness of our results.This three performance metrics are of great significance to the research of link scheduling problems in wireless networks.What's more,the thesis proposes three algorithms on wireless link scheduling problem.We propose a new model,named conflict graph model based on SINR,and the first two algorithms are single slot scheduling algorithm and shortest link scheduling algorithm designed on this new model.They have improved the throughput performance and runtime respectively,compared with the previous algorithms.In the third algorithm,we propose a new notion named interference factor and introduce successive interference cancellation technology.Considering the traffic demands of the links,we design a shortest link scheduling algorithm under layered physical model,which achieves a constant approximation.We prove the correctness and feasibility of the three algorithms by theoretical analysis and simulation experiments,as well as their improvement to the performance of the network.