Confirming that Mathematica 9 can easily solve this system unlike ver. 7 & 8 I'm going to suggest how to deal with it in earlier versions.
- there are many many complex solutions so the restriction of the domain to
Reals is important to get the only one real solution.
- since the system appears to be difficult for
Mathematica 7 & 8 one should consider a simple transformation of the original variables.
There are terms 2 x + y, 6 x + 3 y, 8 x^2 + 4 x y so one can conclude it should be a good idea to introduce a new variable z == 2x + y, now we have :
system1 = { 1 + Sqrt[2 x + y + 1] == 4 (2 x + y)^2 + Sqrt[6 x + 3 y],
(x + 1) Sqrt[2 x^2 - x + 4] + 8 x^2 + 4 x y == 4} /. {y -> z - 2 x}//Simplify
{ 1 + Sqrt[1 + z] == Sqrt[3] Sqrt[z] + 4 z^2,
8 x^2 + (1 + x) Sqrt[4 - x + 2 x^2] + 4 x (-2 x + z) == 4}
and this system appears to be much easier to solve :
Reduce[system1, {x}, Reals]
z == 1/2 && x == 1/2
We can use as well Reduce[system1, {z, x}, Reals] to get the solutions instantly unlike in this case Reduce[system1, {x, z}, Reals] i.e. when the specified variables to be found are in the reversed order.
We can use also Solve eliminating one variable :
Solve[system1, {x}, {z}, Reals]
Solve[system1, {z}, {x}, Reals]
{{x -> 1/2}}
{{z -> 1/2}}
NSolve[system,vars]and postprocess to remove the complex-valued solutions. – Daniel Lichtblau Mar 16 at 20:49