1

I want to solve this system but I can't

Solve[(1 - x^2) (-W - 3 x + Sqrt[6] Sqrt[z])
   == 
   -W^4 x (-1 + x^2) y +
   3 W y (W^3 (-1 + x^2) (-1 + Sqrt[1 - x^2] y) - 
   6 Sqrt[6] (y + Sqrt[1 - x^2] (-1 + z))^2 Sqrt[z] + 
   W^3 (-1 + x^2) z) 
   == 
   9 Sqrt[6] (y + Sqrt[1 - x^2] (-1 + z))^2 (1 - 2 z) Sqrt[z] 
   -1/3 W^3 (-1 + x^2) z (Sqrt[6] W x 
   - 9 (-1 + Sqrt[1 - x^2] y + z)) 
   ==
   3 W (W^3 (-1 + x^2) (-1 + Sqrt[1 - x^2] y) - 
   6 Sqrt[6] (y + Sqrt[1 - x^2] (-1 + z))^2 Sqrt[z] + 
   W^3 (-1 + x^2) z)
   == 0, {x, y, z, W}]
george2079
  • 38,913
  • 1
  • 43
  • 110
milad
  • 21
  • 5
  • it does not appear to be a system ie a list of equations but a single expression with consecutive Equal signs.Consider breaking up the successive equalities into their corresponding parts ie transform x==y==z into {x==y,y==z,x==z} – user42582 Oct 26 '16 at 07:59
  • 1
    @user42582 That's not a problem. It's valid syntax. (BTW it can be expanded using LogicalExpand.) I believe that reason why Solve is so slow that after eliminating all the square roots it ends up with a high order polynomial system of equations with complicated coefficients. It is simply too slow to work with such large expressions. This also means that the solution would look uselessly complicated, so it's not worth computing the exact symbolic result. – Szabolcs Oct 26 '16 at 08:52
  • Evidence for this is also the high memory usage of the kernel when evaluating this Solve. On my machine it keeps hovering around 1 GB (going up and down) after running for a couple of minutes. – Szabolcs Oct 26 '16 at 08:54
  • @Szabolcs you are right , when run this code about all of my RAM and a large part of my CPU used. I am waiting for about 1 hour with any result. – milad Oct 26 '16 at 09:55
  • 1
    I suggest you use NSolve. It also took a long time but it finished and found 967 solutions. I don't know if this solution set is exhaustive. – Szabolcs Oct 26 '16 at 10:10
  • @Szabolcs : Unfortunately, I can't realize NSolve. It crashes my intermediate comp in short time. – user64494 Oct 26 '16 at 10:54
  • @Szabolcs , of the 976 do any have non-zero W? see if you get any solutions if you divide eq 2&4 by W.. ( You must have a better computer than me! ) – george2079 Oct 26 '16 at 15:07
  • NSolve might not recognize that it has an infinite solution set with one and two dimensional components. – Daniel Lichtblau Oct 26 '16 at 16:34
  • Welcome to Mathematica.SE! 1) As you receive help, try to give it too, by answering questions in your area of expertise. 2) Take the tour and check the faqs! 3) When you see good questions and answers, vote them up by clicking the gray triangles, because the credibility of the system is based on the reputation gained by users sharing their knowledge. Also, please remember to accept the answer, if any, that solves your problem, by clicking the checkmark sign! –  Oct 29 '16 at 04:52

1 Answers1

1

Making use ofFindRoot instead of Solve, one obtains by

FindRoot[{sys}, {{x, 0.1} , {y, 0.1}, {z, 0.1}, {W,0.1}}]

{x -> 0.57735, y -> 0.100001, z -> 0.5, W -> 9.21509*10^(-18)}.

Addition. Subsituting W->2 in the system and NSolving it in $x,y,z$, one obtains

NSolve::infsolns: Infinite solution set has dimension at least 1. Returning intersection of solutions with -((121484 x)/178835)-(113492 y)/178835+(171802 z)/178835 == 1.

and

{{x -> 1., y -> -2.64617, z -> 0.}, {x -> -1., y -> -0.505331, z -> 0.}, {x -> -1., y -> 0., z -> 0.33382}, {x -> -1., y -> 0., z -> 0.33382}, {x -> -1., y -> 0., z -> 0.33382}, {x -> 1., y -> 0., z -> 1.74805}, {x -> 1., y -> 0., z -> 1.74805}, {x -> 1., y -> 0., z -> 1.74805}}

Addition 2. By the changes Sqrt[z]==s and Sqrt[1-x^2]==t the system under consideration can be reduced to a polynomial system over Q(Sqrt[6]). The reduced system can be solved in terms of higher degree polynomials via Groebner on powerful comp. It is clear that solution would be huge and useless. I find it monkey business.

user64494
  • 26,149
  • 4
  • 27
  • 56