@tomleslie 

First of all, thank you very much for your attention to my problem and for providing a solution to it.

I think there is a misunderstanding. I meant the proposed solution by @Christopher2222  (and not yours),  which was to use *coeff(A,epsilon)*  to extract all of epsilon's power, doesn't work for integer power of epsilon (I uploaded a screen shot of my Worksheet  for him). This fact was accepted by him in his next post.

****************************************************

> A := d*n0/dt+d*epsilon*n1/dt+d*epsilon^(3/2)*n2/dt+d*epsilon^2*n3/dt+epsilon*d*n0*u1/dx+epsilon^(3/2)*d*n0*u2/dx+d*n0*epsilon^2*u3/dx+d*epsilon^2*n1*u1/dx+d*epsilon^(5/2)*n1*u2/dx;

coeff(A, epsilon^2);
%;
Error, unable to compute coeff
> coeff(A, epsilon);
%;
Error, unable to compute coeff

****************************************************

Anyway, unfortunately I haven't had a chance to check your submitted code yet, but as I said before, your idea is definitely great.

Meanwhile, I also use the Mapel 11 version, which is apparently older than your version, so I hope there is no problem using your code.

Thank you and all the friends who tried to solve my problem.

@tomleslie 

Dear friend,

I could pull out the fractional power of epsilon from the expersion with coeff (), but it does not work for integer power of epsilon!!

Please, check the expersion!!

However, as you say, it seems that introducing a dummy variable is a good idea for integer power of epsilon!

Thanks

@Christopher2222 

Thanks Christopher2222 

Yes, d means differential with respect to x/t

Also, n0, n1,n2,u1,.. are functionn of x/t, but epsilon is a constant!

I tried to send the file, Please see the attached file.

Thanks Christopher2222 5030 

Yes, d means differential with respect to x/t

Also, n0, n1,n2,u1,.. are functionn of x/t, but epsilon is a constant!

I tried to send the file, Please see the attached file.

@Kitonum 

Thanks Sir!

But another question:

why can't I extract the coefficints of $epsilon^2$  or $epsilon^1$ with your proposed response?

see:

> A := d*n0/dt+d*epsilon*n1/dt+d*epsilon^(3/2)*n2/dt+d*epsilon^2*n3/dt+epsilon*d*n0*u1/dx+epsilon^(3/2)*d*n0*u2/dx+d*n0*epsilon^2*u3/dx+d*epsilon^2*n1*u1/dx+d*epsilon^(5/2)*n1*u2/dx; coeff(A, epsilon^2);
%;
Error, unable to compute coeff
> coeff(A, epsilon);
%;
Error, unable to compute coeff
 

@ecterrab 

Thanks you for your attention. Really, I have tried to do that by 'dcoeffs', but it did not work well.

I'll appreciate if you help me.

Please, see the attached file.