
Stochastic representation of solutions of linear PDEs 21
Example 1.6 Digital Neumann options and Az´ema-Yor martingales
Let us consider Φ(x, M) = 1
x≥ζ
(with ζ > 0) and D
0
= {(x, M) | x ≥ g(M)}
with g an increasing function such that g(M ) < M. Equations (1.30)–(1.32)
read
∂
2
x
u(x, M) = 0 g(M) < x < M (1.35)
∂
M
u(x, M)|
x=M
= 0 (1.36)
u(g(M), M) = 1
g(M)≥ζ
(1.37)
Using Equations (1.35) and (1.37) we get
u(x, M) = a(M )(x − g(M)) + 1
g(M)≥ζ
u(X
t
, M
t
) is an example of an Az´ema-Yor’s martingale (see [42] and Exercise
3). Using the remaining equation (1.36), we get
a(M) = αe
R
M
0
g
0
(y)
y−g(y)
dy
−
1
g(M)≥ζ
g
−1
(ζ) − ζ
e
R
M
g
−1
(ζ)
g
0
(y)
y−g(y)
dy
The integration constant α is fixed