Abstract: “In-order” data refer to data that exhibit priority characteristic. To enhance the user quality of experience, the re-ordering delay and decoding delay of in-order data should be minimized. With imperfect feedbacks, the base station can only receive partial user feedbacks and cannot retransmit efficiently. To solve this problem, this paper firstly analyzes the effect of feedback lost on retransmission efficiency, and then proposes three heuristic packets update policies to relieve the negative effect. Extensive simulation results show that the completion time and the total access time of full packets update policy are the best among all proposed policies.

Key words: imperfect feedbacks, in-order data, progressive transmission, multicast, wireless networks

摘要： “有序”数据指的是数据间具有特定优先级关系的一类数据。降低“有序”数据传输时的排序时延和解码时延是提高“有序”数据业务的用户服务体验的关键问题。在非理想反馈条件下，基站仅能正确接收到部分用户反馈信息，导致重传效率降低。针对反馈丢失情况下的包重传问题进行研究，对发送端无法接收到反馈信息的事件进行了分析，提出了三种启发式的包更新策略。大量的仿真实验结果表明，在所提出的三种包更新策略中，全包更新策略对应的包重传方案的完成时间和总接入时间优于其余两种策略。

关键词: 非理想反馈, 有序数据, 渐进式传输, 多播, 无线网络