Research On Instantly Decodable Network Coding With Feedback Loss

Network coding has great advantages to enhance the performance of wireless networks.However,the destination nodes can not decode real-timely,because they need to receive enough encoding information.The IDNC technology chooses corresponding packets to ensure real-time decoding,and it has being widely concerned by scholars.Most previous researches are based on perfect wireless link,but all kinds of fading and interference exist in the practical network.So this thesis mainly researches on IDNC with feedback loss of wireless link,and we do some related work as follows:Firstly,a new network coding model is proposed to establish the relationship between IDNC and RLNC with feedback loss.After constructing of IDNC graph,the sub-generation is proposed,which is built upon optimal IDNC solutions.RLNC is applied in each sub-generation after the partition of sub-generation.Simulation results illustrate that the system performance of different sizes of sub-generation from 1 to IDNCC is between that of IDNC and RLNC.IDNC and RLNC are only two extreme examples with specific sub-generation sizes.Secondly,the purpose is to minimize completion delay and decoding delay of IDNC in multicast network with feedback loss.The expression of average completion delay and decoding delay is obtained after constructing the IDNC graph through MLSS method.Then the clique needed is searched by using the maximum weight vertex search algorithm.Results show that the MLSS algorithm outperforms the blind IDNC graph updating method in other literatures to find the best coding packet.Thirdly,the BCD-FL strategy is proposed to balance completion delay and decoding delay.This paper analyzes this problem by SSP in broadcast network,and the SSP description of the BCD-FL problem in multicast network is obtained by performance expressions that have been reached.Then BCD-FL algorithm is applied to search the maximum clique.Simulation results indicate the proposed BCD-FL algorithm can achieve a good balance of completion delay and decoding delay.