摘要
分析了前馈神经网络火灾探测的基本原理；针对ＢＰ算法的缺陷，提出用遗传算法（ＧＡ）和ＢＰ相结合的一种方法，即：ＢＰ—ＧＡ算法；并把ＢＰ—ＧＡ算法用于火灾探测的模拟，结果表明：前馈神经网络是处理火灾信号的良好方法。
The basic theories of Feedforward Neural Networks For Fire Detection are analyzed. A Method of combining BP Algorithm and Genetic Algorithm (BP-GA) is presented because of BP's limitation. BP-GA Applied in fire detection, the results are manifested: It is a good way to process the fire signal using Feedforward Neural Networks, BP-GA inproves thd networks' convergence.
出处
《火灾科学》
CSCD
1998年第2期21-26,共6页
Fire Safety Science
