@Preben Alsholm You're right. Actually now when I rerun the OPs worksheet (Maple 2017) I'm not getting any errors; it's just slow, Not sure why I thought that.

@Anthrazit  I hadn't thought of the case where the two units didn't have the same zero; that does seem to violate the ideas of quantity calculus. Actually, I am more worried about the case where you can convert 0 K to 0 kg. That's why I used the word "debatable"

Notice that there was a missing space  or * between Pi and (ex6r)^2 so these weren't being multiplied. @Kitonum fixed this up in his answer. The fact that (4)^2 wasn't automatically simplified is a clue that something is not quite as intended.

@isifesai Your solve syntax needs to be

solve({cat(eq, 1 .. 62), seq(a[i] >= 0, i = 1 .. 26), seq(b[i] >= 0, i = 1 .. 12), seq(x[i] >= 0, i = 1 .. 24)}, [seq(x[i], i = 1 .. 24)])

but you have many more equations and constraints than variables, and you don't get any solutions.

@fatemeh1090 The .pdf result has x__2 in, which in the worksheet I did, is expressed in terms of alpha, beta etc. It is not easy to check if they are the same.

@Carl Love That's a nice result. When I get that message, I assume that dsolve has already tried to use solve. So I'm surprised it can be done by hand; I'll try harder with future cases.

The square roots are giving you the problem. You might have better luck formulating the problem in polar coordinates (see SetCoordinates).

I don't get the warning in Maple 2017. Adding the allsolutions option still gives only the two solutions you got, suggesting there are no others.

@baharm31 Sorry, I don't understand what you want. You need to give a simpler example that illustrates what you want.

@tomleslie OP used combine, which converts cos(x)^7 to a sum of cos(n*x) and sin(n*x) type terms, i.e. a fourier series, so there will be no powers like cos(x)^7.

Perhaps:

Round:=(x,digs::nonnegint)-> evalf(round(x*10^digs)/10^digs);

@Kitonum or last line could be add(Dists)/20;

@lcz I haven't used Maplets, but an option like foreground=red can change some things, e.g. in a TextField or Button, though I don't see how to do it for GridLayout.

@Carl Love This leaves "1" as nonlinear, which is not what the OP wanted.

@Carl Love OP's code looks at nops, so would have 3*x nonlinear. But perhaps the ground rule excluded that: "can be any combination of the given variables, that is, x, y, and z", which made me wonder if "y*y" would count.