34 2. FOUR IMPORTANT LINEAR PDE as ε → 0. Hence our sending ε → 0 in (24) yields the formula (25) u(x) = ∂U Φ(y − x) ∂u ∂ν (y) − u(y) ∂Φ ∂ν (y − x) dS(y) − U Φ(y − x)Δu(y) dy. This identity is valid for any point x ∈ U and any function u ∈ C2( ¯ U ). Now formula (25) would permit us to solve for u(x) if we knew the values of Δu within U and the values of u, ∂u/∂ν along ∂U. However, for our application to Poisson’s equation with prescribed boundary values for u, the normal derivative ∂u/∂ν along ∂U is unknown to us. We must therefore somehow modify (25) to remove this term. The idea is now to introduce for fixed x a corrector function φx = φx(y), solving the boundary-value problem (26) Δφx = 0 in U φx = Φ(y − x) on ∂U. Let us apply Green’s formula once more, to compute (27) − U φx(y)Δu(y) dy = ∂U u(y) ∂φx ∂ν (y) − φx(y) ∂u ∂ν (y) dS(y) = ∂U u(y) ∂φx ∂ν (y) − Φ(y − x) ∂u ∂ν (y) dS(y). We introduce next this DEFINITION. Green’s function for the region U is G(x, y) := Φ(y − x) − φx(y) (x, y ∈ U, x = y). Adopting this terminology and adding (27) to (25), we find (28) u(x) = − ∂U u(y) ∂G ∂ν (x, y) dS(y) − U G(x, y)Δu(y) dy (x ∈ U), where ∂G ∂ν (x, y) = DyG(x, y) · ν(y) is the outer normal derivative of G with respect to the variable y. Observe that the term ∂u/∂ν does not appear in equation (28): we introduced the corrector φx precisely to achieve this. Suppose now u ∈ C2( ¯ U ) solves the boundary-value problem (29) −Δu = f in U u = g on ∂U, for given continuous functions f, g. Plugging into (28), we obtain

Purchased from American Mathematical Society for the exclusive use of nofirst nolast (email unknown) Copyright 2010 American Mathematical Society. Duplication prohibited. Please report unauthorized use to cust-serv@ams.org. Thank You! Your purchase supports the AMS' mission, programs, and services for the mathematical community.