Contact linearization of nondegenerate Monge-Ampère equations (Q1006957)

The classical Monge-Ampere equation is the fully nonlinear equation \[ Av_{xx}+2Bv_{xy}+Cv_{yy}+D(v_{xx}v_{yy}-v_{xy}v_{xy})+E=0, \] where \(A=A(x,y)\), \(B=B(x,y)\), \(C=C(x,y)\), \(D=D(x,y)\), and \(E=E(x,y)\). This class of equations is closed under the pseudogroup \(G\) of contact transformations. One says that such a problem can be linearized if there is a change of coordinates which transforms this to the (quasi) linear equation \[ av_{xx}+2bv_{xy}+cv_{yy}+rv_x+sv_y+kv+w=0 \] where again \(a=a(x,y)\), \(b=b(x,y)\), \(c=c(x,y)\), \(r=d(x,y)\), \(s=s(x,y)\), \(k=k(x,y)\), and \(w=w(x,y)\) are functions of the coordinates. The authors solve the problem of linearization (with respect to the pseudo-group \(G\)) of regular elliptic and hyperbolic Monge-Ampere equations by constructing invariants of these equations. The linearization process is constructive and requires only first integrals of completely integrable distributions. It involves constructing differential tensor invariants in this context.