Assume equation (2.32) can be written in the form
F(x,y,z,p,q) = 0, (2.32)
∂x∂z=f(x,y,z,∂y∂z) (2.41)
(for example, in a neighbourhood of a point where it holds ∂p∂F=0.) Then using the method of variation of constants, prove that all the solutions of (2.32) can be obtained from the complete solution.
Hint. Firstly take a = a( x, y) and b = b( x, y) in the complete solution
z = g(x,y,z,a,b).
Then, by the definition of the complete solution, replacing z, a and b with the functions g(x,y,z,a,b), ∂x∂g(x,y,z,a,b) and ∂y∂g(x,y,z,a,b) respectively, one gets an identity in (2.41) for arbitrary functions a and b.
Conversely, a function z = g(x, y, z, a, b) is a solution of (2.41) only if
∂x∂z=∂x∂g and ∂y∂z=∂y∂g.
Show that this will hold only if
∂a∂g(x,y,a,b)∂x∂a+∂b∂g(x,y,a,b)∂x∂b=0∂a∂g(x,y,a,b)∂y∂a+∂b∂g(x,y,a,b)∂y∂b=0. (2.42)
Putting ∂a∂g(x,y,a,b)=0 and ∂b∂g(x,y,a,b)=0, and, if possible, eliminating a and b from equations (2.42), we obtain the singular solution. Note that it depends neither from arbitrary constants nor from arbitrary functions.
Next, assume it holds ∂(x,y)∂(a,b)=0, where
∂(x,y)∂(a,b)=∣∣∣∣∣∣∂x∂a∂x∂b∂y∂a∂y∂b∣∣∣∣∣∣,
and let a and b be some non constant functions. Then the functions a and bare not functionally independent, i.e., there exists a differentiable function w such that b = w(a). This leads to the general solution of (2.41).
Finally, check that any solution of (2.41) reduces to one of the following three ones: singular, complete or general.