matlabFunction on symbolic experssion and the Error using symengine

Hello

I have a mix double integral and partial differential expression and I want to solve it by symbolic math in matlab.

My code below written to solve the equation. But matlabFunction() does not work on it and gives me an error.

I would be appreciated if anybody could help me. I tried the numerical way to solve this expression but I could not find the right result. The only way that I think work is using symbolic math. Any idea and help is appreciated.

Here is my code:

syms t z
w=1.11;
T=1.2;
zita=0.5;
a=1;
b=1;
d=0;
%%
teta=exp(-((1-t/T)*(t/T))^(-w));
phi=int(teta)/int(teta,0,T);
Y1=[1;1]+[1;2]*phi;
Y2=[2;1]+[1;1]*phi;
eta0=Y1+(z-zita)*Y2;
F=[(1/a)*(diff(eta0(1,1),t)-eta0(2,1)*eta0(1,1)); ...
       (1/b)*(diff(eta0(2,1),t)-d*eta0(2,1))];

int1=int(F,z)-int(F,z,0,zita);
etaAux=int(int1,z)-int(int1,z,0,zita);
eta=eta0+etaAux;
etaD=diff(eta,z);
fun=matlabFunction(etaD);

After running the code, I got this error:

Error using symengine
Unable to evaluate to Boolean.

Error in sym/matlabFunction>mup2mat (line 404)
res = mupadmex('symobj::generateMATLAB',r.s,ano,spa,0);

Error in sym/matlabFunction>mup2matcell (line 374)
    r = mup2mat(c{1},true,sparseMat);

Error in sym/matlabFunction (line 188)
    body = mup2matcell(funs, opts.Sparse);

Could you please help me with this issue?

etaD is 2by1 and matlabFunction() works on etaD(2,1) but does not work on etaD(1,1) and gives me the same error as above.

[Walter Roberson]

It has something to do with the square of an integral. It doesn't matter if it is numerator or denominator but it does matter that it is squared

Answer 1

[Walter Roberson]

It appears to be a bug. I might look into the internal code when I am more awake, but that code is difficult to read and debug.

In the meantime, there is a workaround. It takes a bit of time to compute it, but it gets there.

The workaround is instead of

matlabFunction(etaD)

use

matlabFunction(rewrite(etaD, 'sincos'))

You will have to wait a bit for it to compute the rewrite()

Do not simplify() the results of the rewrite() or else you get back to code that has the problem

[Amin NZH]

Thank you so much for your answer.

It works and does not have any error but, the result is not true.

I compare it with matlabFunction(etaD(2,1)) ( which it works):

fun1=matlabFunction(etaD(2,1));
fun1(0.5,0.5)
ans=
-0.55

fun2=matlabFunction(rewrite(etaD(2,1),'sincos'));
fun2(0.5,0.5)
ans=
NaN

I would be really appreciated if you help me and take a look at the code any time you are free.

Many thanks in advance

[Walter Roberson]

Debugging this further, I see that it occurs under the circumstance that you have an int() to a power, and the int() does not have a closed form, and the int() expression can be evaluated to a floating point number.

I am experimenting with using vpa() on etaD, but something odd seems to be happening.

[Walter Roberson]

Yes, you can matlabFunction(vpa(etaD))

I will file a bug report with Mathworks.

[Amin NZH]

IT WORKS!

Thank you so much for your help and time.

I appreciate your kind guidance.

Best regards