论文部分内容阅读
研究遗传神经网络算法在异步转移模式（ＡＴＭ）链路容量分配中的应用。利用神经网络预测器，借助足够的学习样本，自动学习用户信息和网络性能的未知关系，预测网络性能，并用遗传算法进行优化分配，解决了ＡＴＭ链路容量分配中的困难。分别采用链路利用率和呼损率两种优化目标进行计算机模拟。结果表明，该方法能正确完成容量分配，与其它方法相比，不需对网络进行复杂的数学分析，控制简单、易行。
Research on Genetic Neural Network Algorithm for Asynchronous Transfer Mode (ATM) Link Capacity Allocation. With the help of neural network predictor, enough learning samples are used to automatically learn the unknown relationship between user information and network performance to predict the network performance. Genetic algorithms are used to optimize the allocation, which solves the difficulties in capacity allocation of ATM links. The two optimization objectives of link utilization and call loss rate are respectively used for computer simulation. The result shows that this method can finish the capacity allocation correctly. Compared with other methods, this method does not need complicated mathematical analysis of the network, and the control is simple and easy.