Abstract: Concerning the problem of being easy to generate congestion during data aggregation in power distribution and utilization communication network, a novel hybrid edge-weighted traffic scheduling and routing algorithm was proposed. Firstly, the hierarchical node model was established according to the number of hops. Then, the priorities of power distribution and utilization services and node congestion levels were divided. Finally, the edge weights were calculated based on the comprehensive index of hop number, traffic load rate and link utilization ratio. The nodes of traffic scheduling were needed for routing selection according to the improved Dijkstra algorithm, and the severe congestion nodes were also scheduled in accordance with the priorities of power distribution and utilization services. Compared with Shortest Path Fast (SPF) algorithm and Greedy Backpressure Routing Algorithm (GBRA), when the data generation rate is 80 kb/s, the packet loss rate of emergency service by using the proposed algorithm is reduced by 81.3% and 67.7% respectively, and the packet loss rate of key service is reduced by 79% and 63.8% respectively. The simulation results show that, the proposed algorithm can effectively alleviate network congestion, improve the effective throughput of network, reduce the end-to-end delay of network and the packet loss rate of high priority service.

Key words: power communication network, improved Dijkstra algorithm, traffic scheduling, composite edge weight, service priority

摘要： 针对配用电通信网中数据汇聚易产生拥塞的问题，提出了一种复合边权值流量调度路由算法。首先，依据跳数建立节点分层模型；然后，划分配用电业务优先级和节点拥塞等级；最后，以跳数、流量负载率和链路利用率为综合指标计算边权值，对需要流量调度的节点根据改进的Dijkstra算法进行路由选择，同时对重度拥塞节点按照配用电业务优先级进行调度。与最短路径（SPF）算法和贪婪背压算法（GBRA）相比，在数据生成率为80 kb/s时，所提算法紧急型业务丢包率分别减少了81.3%和67.7%，关键型业务丢包率分别减少了79%和63.8%。仿真结果表明，所提算法能有效缓解网络拥塞，提高网络有效吞吐量，降低网络端到端时延和高优先级业务的丢包率。

关键词: 电力通信网, 改进的Dijkstra算法, 流量调度, 复合边权值, 业务优先级