@emendes 

1. [f](...) is equivalent to [f(...)] for any name f (stock command or otherwise). Using the former helps to reduce long strings of closing brackets at the end of expressions, which is a purely stylistic concern.
2. _ is a name, just like A.
3. A^+ is the (matrix, array, vector) transpose of A.
4. The first procedure is threadsafe, or either seq could be replaced by Threads:-Seq. The second procedure may not be threadsafe because of the transpose (which calls LinearAlgebra:-Transpose).
5. As far as I know, there are no formal restrictions on code that can be run with Grid, provided that you have the memory---there is no concept "gridsafe". I don't understand why Threads is not working after Grid has finished; that is a bug. You could save data, then restart, read in the data, then use Threads.

Having such a large expression entered in 2D Input makes subsequent input excruciatingly slow to enter, even if that subsequent input is done in 1D.

@Tamour This is not a place to promote illegal uses of intellectual property (IP).

@itsme Yes, that's a good improvement.

Please attach the worksheet by using the green uparrow on the MaplePrimes editor toolbar.

@acer You're right, and I think that subs(n= nn, ...) is the best solution (vote up). I just wanted to point out the existence of infix %assuming, which is new to Maple 2020 I believe.

However, the weirdness in this case arises because assuming treats Sum ​​​​​​differently than other unevaluated functions (which you're probably aware of, but the other correspondents are not), so it makes some sense to avoid using Sum in the first place. Compare with

restart:
e:= SUM((-1)^n-1, n= 1..infinity):
simplify(op(1,e)) assuming n::even;
                               0

@Scot Gould I was not expecting my idea to work in the case of a new, unsaved worksheet.

@acer I meant this:

restart:
e:= %Sum((-1)^n-1, n= 1..infinity);
value(simplify(op(1, e) %assuming n::even));
 

@krobe8 You could put

currentdir(interface(worksheetdir));

in your initialization file.

@acer A recent upgrade, perhaps 2020, allows %assuming to be used in infix form. There are some examples at ?assuming.

@perr7 So why have you ignored the thickness option?

@perr7 The difference between your originally posted plot and the one immediately above cannot be due to Kitonum's suggestion. There's either a difference in your configuration (possibly the Digits setting) or a difference in your P_QM.

@Anthrazit I don't understand all the details of what's going on, but the following transcript proves that with(Units) does replace (strictly speaking, rebind) the default max with Units:-Simple:-max and that the first sentence of my Answer is not wrong (no offense taken):

restart:
with(Units):
Automatically loading the Units[Simple] subpackage

showstat(:-max);
Error, (in showstat) cannot debug built-in functions

showstat(max);
Units:-Simple:-max := proc()
    [...procedure code redacted, but you can view it yourself...]
end proc

@Kitonum The thickness option can be used for surfaces drawn with style= line. In this case, I think that it makes a big improvement. This is with thickness= 4:

I suspect that in your first differential equation, the second Q[1] is supposed to be Q[2]. Please confirm.