论文部分内容阅读
为了改进亏空公平优先队列算法,提出一种成比例带宽分配算法。基站将服务连接的带宽请求按照优先级排队,根据服务质量参数先分配部分带宽,超出的请求基于子站总请求带宽的大小成比例分配给服务连接。仿真采用2维离散时间Markov调制Poisson过程的模型产生实时轮询服务和非实时轮询服务连接的数据源。与亏空公平优先队列算法相比,该算法不仅满足各类服务连接的服务质量要求,而且改善子站总的吞吐量和流量突发增加的问题。该算法还提高了子站间及不同优先级服务连接间的公平性。
In order to improve the deficit fairness priority queue algorithm, a proportional bandwidth allocation algorithm is proposed. The base station queues the bandwidth requests for service connection according to priorities, and allocates a part of bandwidth according to the service quality parameters. The excess requests are allocated proportionally to service connections based on the total requested bandwidth of the sub-stations. Simulation Models using 2-dimensional discrete-time Markov-modulated Poisson processes generate data sources for both real-time polling services and non-real-time polling service connections. Compared with the unfair priority queue algorithm, this algorithm not only meets the service quality requirements of various service connections, but also improves the total throughput of sub-stations and the sudden increase of traffic. The algorithm also improves the fairness between sub-stations and between different priority service connections.