基于ADAFSA-BP的概率积分法预计参数研究

Parameter prediction of probability integral method based on ADAFSA-BP

在研究人工鱼群算法缺点的基础上,为了提高概率积分法预计参数的准确性,提出改进鱼群算法(ADAFSA),建立ADAFSA-BP预计模型。以我国40个工作面的地质采矿条件为训练集和测试集,将ADAFSA-BP模型与BP模型、GA-BP模型进行对比分析。结果表明:ADAFSA-BP模型对参数的预测平均相对误差均低于10%,满足精度要求,同时在小数据量学习方面ADAFSA-BP模型优于BP模型和GA-BP模型。

On the basis of studying the shortcomings of artificial fish school algorithm,in order to improve the accuracy of probability integral method,an improved fish school algorithm(ADAFSA)is proposed and ADAFSA-BP prediction model is established.Taking the geological and mining conditions of 40 working faces in China as the training set and test set,the ADAFSA-BP model is compared with BP model and GA-BP model.The results show that:the average relative error of ADAFSA-BP model is less than 10%,which meets the accuracy requirements.At the same time,ADAFSA-BP model is better than BP model and GA-BP model in small amount of data learning.