Maximum mu reached training closed-loop Narx net

Question

Luis Ignacio Ruiz 2016-3-9

0
链接

此问题的直接链接

https://ww2.mathworks.cn/matlabcentral/answers/272334-maximum-mu-reached-training-closed-loop-narx-net

Hi, i´m trying to train a closed loop Narx net with the following code, but it just doesn´t iterate because every time it stops at the first iteration saying "Maximum mu Reached". I tried canging the maximun mu and mu_inc but nothing seems to change.

How can i make it train properly?

Thank you beforehand

clc;
clear all;
close all;
%--------------------------------------------------------------------------------------
load('Datos_entren_Senal_2');
% load('datosNarX');
% load('datosNarX2');
% load('datosNarX3');
load('NarX55_18cl');
x = con2seq(EntradasNarX'); 
t = con2seq(SalidasNarX');%Se introducen los datos de entrada y salida correspondientes
net.trainParam.epochs=1000;
% net.trainParam.min_grad=1e-10;
% net.trainParam.show=1;
% net.trainParam.goal = 1e-3;
% net.trainParam.mu_max=1e70;
net.trainParam.mu_inc=2;
net.trainParam.max_fail=18;
[X,Xi,Ai,T] = preparets(net,x,{},t); %se preparan los datos y la red para ser entrenada
% net=adapt(net,x,t);
perfc_f=2;
for i=1:30   
%Datos 1
[net tr y e] = train(net,X,T,Xi,Ai); %se entrena la red con los datos preparados
% view(net); %visualiza la red
Yc = net(X,Xi,Ai); %se obtienen las salidas del entrenamiento
perfc = perform(net,Yc,T);
if perfc<perfc_f
    perfc_f=perfc;
end
end