0
$\begingroup$

I have this system of six equations for $\{x,y,z,u,w,v\}$.

sixeqs = {I (-x + v) + u == (α x)/2 + y    , 
   z + I  (w + v) == u    , 
   1/2 (-2 I + α) x + y + I w ==  z, 
(2 I + α) x +  2 E^(-I (2 β π - β γ +  t)) ((2 π- I - γ) w + z) == 2 y    ,   
 E^(-I (β (π - γ) + t)) ((-2 π - I + γ) w - z +  E^(2 I β π) ((-I + γ) v + u)) ==  0,
 (I - α/2) x + y == E^(I (β γ - t)) ((I + γ) v + u)  };

I want to find a non-zero solution for $\{x,y,z,u,w,v\}$ if it exists; I use Solve and the assumptions I have, but it gives only the trivial solution $\{x,y,z,u,w,v\}=\{0,0,0,0,0,0\}$.

Assuming[{    α > 0 &&α ∈   Reals  &&  γ> 0 &&  γ ∈ Reals   && t ∈ Reals && β ∈ Reals     }, 
 Simplify[ Solve[ sys, {x, y, z, w, u, v}]]]
(*{{x -> 0, y -> 0, z -> 0, w -> 0, u -> 0, v -> 0}}*)

Then, when I replace Solve with Reduce, it takes a lot of running time.

My question:

How can I obtain nonzero solutions for $\{x,y,z,u,v,w\}$?

$\endgroup$

1 Answer 1

4
$\begingroup$

Your six equations seem to be linear and homogenous M.{x, y, z, u, v, w}==0:

eqn = sixeqs /. Equal -> Subtract;
M = D[eqn, {{x, y, z, u, v, w}}];
M . {x, y, z, u, v, w} == eqn // Simplify(*True*)

Condition for nontrivial solution is Det[M]==0

tmp=Simplify[Det[M], {\[Alpha] > 0 && \[Alpha] \[Element] Reals && \[Gamma] >0 && \[Gamma] \[Element] Reals && t \[Element] Reals && \[Beta] \[Element] Reals}] 

    (*8 E^(-I (3 t + 3 \[Pi] \[Beta] - \[Beta] \[Gamma])) (E^(
    I (t + \[Beta] \[Gamma])) (-2 I + \[Alpha]) + 
   E^(I (t + \[Beta] (4 \[Pi] + \[Gamma]))) (2 I + \[Alpha]) - 
   2 E^(I (t + \[Beta] (2 \[Pi] + \[Gamma]))) (2 \[Pi] + \[Alpha]) + 
   E^(2 I \[Beta] (\[Pi] + \[Gamma])) (-2 I + 2 \[Pi] - \[Gamma]) + 
   E^(2 I (t + \[Pi] \[Beta])) (2 I + 2 \[Pi] - \[Gamma]) + 
   E^(2 I (t + 2 \[Pi] \[Beta])) (-2 I + \[Gamma]) + 
   E^(2 I \[Beta] \[Gamma]) (2 I + \[Gamma])) *)
   

Addendum*

Using this condition the nontrivial solution follows:

cond = Solve[tmp == 0, \[Alpha]][[1]];
sol=Solve[(M /. cond) . {x, y, z, u, v, w} == 0, {x, y, z, w, u,v}];
Simplify[%, {\[Alpha] > 0 && \[Alpha] \[Element] Reals && \[Gamma] >0 && \[Gamma] \[Element] Reals && t \[Element] Reals && \[Beta] \[Element] Reals}] 

    (*{{y -> (1/(2 (-1 + E^(2 I \[Pi] \[Beta]))^2))
   E^(-I (t + \[Beta] \[Gamma]))
     x (E^(2 I \[Beta] (\[Pi] + \[Gamma])) (-2 I + 
         2 \[Pi] - \[Gamma]) - 
      E^(2 I (t + 2 \[Pi] \[Beta])) (-2 I + \[Gamma]) + 
      E^(2 I \[Beta] \[Gamma]) (2 I + \[Gamma]) + 
      E^(2 I (t + \[Pi] \[Beta])) (-2 I - 2 \[Pi] + \[Gamma])), 
  z -> (1/((-1 + E^(2 I \[Pi] \[Beta]))^2))
   E^(I \[Beta] (2 \[Pi] - \[Gamma]))
     x (-I E^(I \[Beta] (2 \[Pi] + \[Gamma])) + 
      E^(I \[Beta] \[Gamma]) (I + 2 \[Pi]) - 
      E^(I (t + 2 \[Pi] \[Beta])) (-I + \[Gamma]) + 
      E^(I t) (-I - 2 \[Pi] + \[Gamma])), 
  w -> (E^(2 I \[Pi] \[Beta]) (1 - E^(
      I (t - \[Beta] \[Gamma]))) x)/(-1 + E^(2 I \[Pi] \[Beta])), 
  u -> (1/((-1 + E^(2 I \[Pi] \[Beta]))^2))
   E^(-I \[Beta] \[Gamma])
     x (I E^(I \[Beta] \[Gamma]) + 
      E^(I \[Beta] (2 \[Pi] + \[Gamma])) (-I + 2 \[Pi]) - 
      E^(I (t + 4 \[Pi] \[Beta])) (-I + \[Gamma]) + 
      E^(I (t + 2 \[Pi] \[Beta])) (-I - 2 \[Pi] + \[Gamma])), 
  v -> (x - E^(I (t + 2 \[Pi] \[Beta] - \[Beta] \[Gamma])) x)/(
   1 - E^(2 I \[Pi] \[Beta]))}}*)
$\endgroup$
9
  • 1
    $\begingroup$ Your question reads as "How can I make sure if the given system of equations has a non-zero solution?" and I gave you the condition! $\endgroup$ Commented Mar 31, 2022 at 10:12
  • $\begingroup$ It was my bad in writing my question. Do you have any idea for calculating non-zero solutions of $\{x,y,z,u,v,w\}$? @Ulrich Neumann $\endgroup$
    – sara96
    Commented Mar 31, 2022 at 10:16
  • 1
    $\begingroup$ @sara96 See my modified answer... $\endgroup$ Commented Mar 31, 2022 at 10:50
  • 1
    $\begingroup$ The solution shows ` {y,z,u,v,w} ` all depend on x. That's the message, warning tried to give. $\endgroup$ Commented Mar 31, 2022 at 11:08
  • 1
    $\begingroup$ No it shouldn't: Your equation has only nontrivial solution if the condition is fullfilled! $\endgroup$ Commented Mar 31, 2022 at 12:29

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.