论文部分内容阅读
为提高智能配电通信业务的服务质量,根据智能配电网对通信技术的要求,提出一种基于动态模糊神经网络(DFNN)的智能配电异构无线网络准入控制算法。在智能配电网络的异构准入控制模型中构建神经网络系统,以网络的接入阻塞率差作为系统参数强化学习的目标,对网络的负载均衡具有较好的动态适应性。神经网络系统在输入层较多时容易产生太多规则而影响决策结果,而DFNN通过计算当前系统规则的完备性,动态添加规则,并通过计算所有规则的重要性,动态删除规则,使得系统的规则有效而不冗余。仿真结果表明,该方法较多接入选择算法(MLB)明显降低了网络的接入阻塞率,相对于模糊神经网络算法(FNN)而言简化了系统结构,突出了规则的重要性,具有较低的接入阻塞率和更好的均衡效果。
In order to improve the service quality of intelligent distribution and communication services, a distributed intelligent network access control algorithm based on dynamic fuzzy neural network (DFNN) is proposed according to the requirement of communication technology in intelligent distribution network. In the heterogeneous admission control model of intelligent distribution network, a neural network system is constructed. With the network access blocking rate difference as the target of system parameter reinforcement learning, it has good dynamic adaptability to the network load balancing. The neural network system tends to produce too many rules when the input layer is too large, which will affect the decision-making result. DFNN dynamically adds rules by calculating the completeness of the current system rules and dynamically deleting them by calculating the importance of all the rules. Effective without redundancy. Simulation results show that this method has more access selection algorithm (MLB) significantly reduces the access blocking rate of the network, simplifies the system structure compared to the fuzzy neural network algorithm (FNN), highlights the importance of the rules, has more Low access blocking rate and better balance.