By factorizing, yes. First note that p(x,y,z) = (x-y+z)2 + x2 + x2 + z2. We have written p as the sum of four squares. Suppose that not all of x,y,z are zero. If at least one of x or z is nonzero, then at least one of the last two squares is positive. If x and z are both zero, then y must be nonzero, so that the first square is positive.
Dorsa R.
asked 03/17/20Can we prove that this equation is positive or negative definite by factorizating it ?
p(x,y,z) = 3x2 + y2 +2z2 - 2xy +2xz - 2yz
2 Answers By Expert Tutors

Matthew S. answered 03/17/20
PhD in Mathematics with extensive experience teaching Calculus
I get p(x, y, z) = (x, y, z)A(x, y, z)T where
A = [3 -1 1]
[-1 1 -1]
[1 -1 2]
There's a theorem that says the quadratic form p is positive definite if and only if the principal minors of A all have positive determinants. The principal minors are:
S1 = the 1,1 entry of A; determinant is 3 > 0
S2 = [3 -1] (i.e., the 2x2 matrix in the upper left-hand corner); determinant of S2 is 3 - 1 = 2 > 0
[-1 1]
S3 = the matrix A itself. det(A) = 3(2 - 1) - (-1) (-2 + 1) + 1(0) = 3 - 1 > 0
Therefore p is positive definite

Matthew S.
I'm not sure what you mean by "square like". I tried calculating the characteristic polynomial of matrix A to see if it could be factored. I didn't succeed in finding a rational root... I may have messed up the calculation.03/17/20
Dorsa R.
that's fine thanks for your effort03/18/20
Still looking for help? Get the right answer, fast.
Get a free answer to a quick problem.
Most questions answered within 4 hours.
OR
Choose an expert and meet online. No packages or subscriptions, pay only for the time you need.
Dorsa R.
thanks for the answer but can we make it square like and get the same answer ?03/17/20