Does Maple have any faciltiy to find any solution to two equations in  3 variables subject to constraint that each equation is not zero?

Here is an example. I like to find any values of t1,t2,t3 such that alpha and beta are not zero. Any choice of such t's will do. But not all t can be zero ofcourse. All the t's are linear.

restart;
eq1:= alpha = -2*t1 - 4*t2 - 2*t3;
eq2:= beta = t1 + 2*t2 + t3;

By inspection we see that choice t2=0,t3=0,t1=1 works since

subs([t2=0,t3=0,t1=1],[eq1,eq2])

There are many other choices (infinite). I just need to find one combination of t_i in the integers.

I could write a loop and keep trying different values of t1,t2,t3 until I find such choise ofcourse. But I was wondering if there is better way to do this in Maple. Solve does not work on this ofcourse.

solve([eq1,eq2,eq3],[t1,t2,t3]) assuming alpha<>0,beta<>0

And setting up Ax=b does not work

restart;
A:=Matrix([[-2,-4,-2],[1,2,1]]);
b:=Vector([alpha,beta]);
LinearAlgebra:-LinearSolve(A,b)

Error, (in LinearAlgebra:-LinearSolve) inconsistent system
 

However, one thing I could do, is pick random alpha,beta values, and then solve for t_i, as in

restart;
A:=Matrix([[-2,-4,-2],[1,2,1]]);
b:=Vector([-2,1]);
LinearAlgebra:-LinearSolve(A,b)

And now I am able to find a solution I want. The problem with this method, is that if I pick wrong values of alpha,beta, I can also get no solution. For example, if I guessed alpha=1,beta=1 I get

restart;
A:=Matrix([[-2,-4,-2],[1,2,1]]);
b:=Vector([1,1]);
LinearAlgebra:-LinearSolve(A,b)

Error, (in LinearAlgebra:-LinearSolve) inconsistent system
 

So one option I could try, is pick random values of alpha,beta, and call LinearSolve until I get a choice which works?  i.e. one which does not give inconsistent system.

Thanks for any better suggestions.

This is the context: I call LinearAlgebra:-LinearSolve to solve system of linear equations. If there are infinite solutions, then Maple return the solution that has free variables, calling them using  _t[n] where n can change each time and be different. It can also be _t0[n] and _t1[n] etc.. but these all are indexed variables.

I want to obtain a list of all these free variables in the resulting solution vector, so that later I can assign them some values.

But I do not know what n can be. And how many of them there are. There could be _t[3] and _t[4] for example in the same solution vector.

Here is what I do now, where I just check for indexed type. This seems to work, since only _t[n] should be in the resulting solution if any.

restart;
A:=Matrix([[0,1,-1],[-1,-2,2],[-1,-2,2]]):
v:=Vector([-1,1,1]):
sol := LinearAlgebra:-LinearSolve(A,v);

map(x->select(type,x,'indexed'),convert(sol,set))

I do not know how to tell it to look for all something that starts with  _t and also indexed type.

Is there a better way to obtain list of all _t[n] that LinearAlgebra:-LinearSolve could return? Will LinearAlgebra always return the free variables as indexed variables so I am sure the above will always work?  Do you see a problem with the above solution?

It is annoying that some functions in Maple wants input like (a,b,c,.....) which represents numbers, and I am not able to find how to use such a function, because the list of numbers I have are in a list.

For example, ilcm and igcd.  This is not a good design. The input should have been a list or set, or vector, etc.... 

I wanted to find least common multiplier of a list of numbers. These numbers are allready in a list, since this is a result of a computation done earlier. Now I want to find ilcm of them. 

How to use ilcm in this case? How to unpack them to make ilcm happy? In Mathematica there a special function to do this, called Sequence, which takes a list and unpack it to call function. 

But I am not able to find one in Maple. There is no convert(list,exprseq). 

Here is a MWE

v:=[2,4,7];
ilcm( v ); #does not work, since ilcm does not accept a list

The variable v above has to be in a list (or set, or vector). This is result from another computation. This is all done non-interactive. 

So I am looking for some magic function to use it like this

v:=[2,4,7];
ilcm( convert_to_expression_sequence(v) )

Is there a way to unpack or convert list to expression sequence, so I can use ilcm?

This is my attempt. But I suspect there is a build-in way in Maple to do this which I have not found yet.

restart;

convert_list_to_exprsequence:=proc(L::list)::exprseq;
local r:=NULL,item;

for item in L do
    r:=r,item;
od;

return r;
end proc:

v:=[2,4,7];
ilcm(convert_list_to_exprsequence(v)); #now it works

                  #28

Maple 2020.1

For entering system of linear ode's, from textbook, such as this example

I can either use x[1],x[2]  or x__1 and x__2.

Maple dsolve works with both. The Latex generated is also exactly the same Which is x_{1} and x_{2}.

The only minor difference I see is on the screen in the worksheet (I only use 1D), where the subscript looks a little nicer in the case of x[1] vs. x__1

The subscript in the case of x[1] is upright while the one in x__1 is italic. I like the first one more.

Other than this cosmetic difference, Are there are subtle sematic differences between the two forms in terms of computation? I know x[1] create table x. OK.  While x__1 is a separate symbol. No table is created.,

Which would be  a safer/better option to use in this case?   Or is it just a matter of choice?

sys:=[ diff(x[1](t),t)=x[1](t)+x[2](t),diff(x[2](t),t)=x[1](t)+3*x[2](t)];
sys:=[ diff(x__1(t),t)=x__1(t)+x__2(t),diff(x__2(t),t)=x__1(t)+3*x__2(t)];

In some way, using x__1(t) might be a little easier to read on the eye than x[1](t) since there is only one type of paranthesis involved.

ps. I tried to search for __ in Maple help to read more about it, but ?__ do not show it. I am sure __ is in the help under some other name.

For display only purposes, to get the Latex looking better, I replace the constant of integrations Maple using which is _C1, _C2., etc... to c[1],c[2],..... 

This is something I do at the very end of computation, just before I get the Latex of the expression.

The way I do this now, is not good. .And I am asking if there is a better way to do this replacement. In Mathematica, I can do this using a patterm where I tell it to replace  C?  by c[?] where ? means anything.

Is it possible to do something like this in Maple? Here is an example

sol:=dsolve(diff(x(t),t$3) = x(t));

This gives solution as

And just before ask for the Latex of this, I change it to 

This is done using a function. Please see worksheet below.

The problem with the current solution is that I do not know how many _C there are, so I assumed 10 as worst case, which works for my case. But I prefer a more general solution, may be using pattern in the subs command?

Download q.mw

And now I get the latex of tmp in the above, instead of sol which looks better.