The Setting
Moment coordinates use two copies of . The source, with coordinates , carries a smooth function and the measure ; the target, with coordinates , carries a smooth potential and the measure for a constant . The two sides are tied, in terms of the gradient and the Hessian of , by the Monge-Ampère equation (1) with a constant and nowhere zero, and by the gradient map , which sends to and is a homeomorphism of with smooth inverse.
On the target side the Hessian of , carried over by , supplies the coefficients of a second-order operator, the moment operator defined below. Against , is computed in two ways, by integration by parts and as the weighted divergence of a vector field, and the second computation is repeated one order up, for a field built from the second derivatives of a function.
Operators in Moment Coordinates
Derivatives are taken in coordinates: is the directional derivative of along the standard basis vector of , where . The gradient is the vector of the ✦, and the Hessian is the matrix with entries ✦.
The moment matrix is the field of matrices on the target given by the Hessian of at the source point of , (2)✦ and in the same way a function on the source is read in target coordinates as ✦.
For a field of matrices on , the operator sends a smooth function to the function (3)✦ and the gradient form of takes two smooth functions and to the function ✦. For , the operator is the moment operator, written , and .✦
The flux of a smooth function is the vector field on the target with components .✦
The weighted adjoint of the directional derivative for the weight is the operator ✦, and the weighted divergence of a vector field on the target is (4)✦ For a vector field on the source, the weighted divergence for the weight is given by (4) with in place of .
The moment coefficients of a smooth function on are the entries of the matrix field on the source given by (5)✦ For each index , the moment divergence of is the function on the source given by the weighted divergence of the -th row of , (6) The vector field on the source collects these functions.
The Claim
Theorem 1. Let and be smooth functions on such that vanishes nowhere and the Monge-Ampère equation (1) holds for a constant , let the gradient map be a homeomorphism of with smooth inverse, and let be a smooth function on . If has compact support, then for every smooth function and every , (7)✦ For every smooth , compactly supported or not, is the weighted divergence of the flux of , (8)✦ the field read in target coordinates is the flux of , (9)✦ and its weighted divergence is (10)✦ The identity (9) holds without the Monge-Ampère equation and without the condition on .
What Is Checked
The four identities are stated over smooth functions , , and on , with nowhere zero, the Monge-Ampère equation (1) and the gradient map a homeomorphism with smooth inverse, and over the moment matrix (2), the operator and gradient form of (3), the flux , the weighted divergence (4) for the weights and , the measure and the field of (6).