| 摘要: |
| 本文证明在ＰＬＮ网络中适当给定广义Ａ－学习律，可使网络满足如下条件：（１）所有训练样本都是稳定状态；（２）每个稳定状态具有最大的吸引域；（３）具有很快的收敛速度．由此可见，将这种ＰＬＮ网络作为联想记忆器是很理想的． |
| 关键词: PLN网络，训练样本，稳定状态，吸引域 |
| DOI： |
| 分类号: |
| 基金项目:国家攀登计划,863高科技支持 |
|
| THE QUANTITATIVE ANALYSIS OF THE ATTRACTIVE REGION OF THE PLN NETWORK |
|
Zhang Ling,Zhang Bo
|
| Abstract: |
| This paper shows that when a generalized A-learning rule of a PLN net is given properly, the net can satisfy the following conditions. (1) All training samples are itsstable states. (2) Each stable state has a maximal attractive region. (3) The net has veryfast convergence speed. Therefore, as an associative memory the PLN network is an idealone. |
| Key words: PLN network, training sample, stable state, attractive region. |
