
184 Nonlinear Optimal Control Theory
The second equation in (6.3.28) therefore becomes
dq
dt
= −A(t)
t
q, (6.6.8)
which is the same as (6.6.5).
Let λ be a solution of (6.6.8) satisfying the initial condition
λ(t
0
) = η. (6.6.9)
Let Ψ denote the fundamental matrix solution of (6.6.8), which satisfies
Ψ(t
0
) = I. Thus,
Ψ
′
(t) = −A(t)
t
Ψ(t) Ψ(t
0
) = I.
Then
λ(t) = Ψ(t)η. (6.6.10)
By (6.6.6), Eq. (6.6.10) can also be written as
λ(t) = (Φ
−1
)
t
(t)η.
Assumption 6.6.3. (i) The functions a
0
and
b
h and the matrix A are C
(1)
on I
0
.
(ii) The end point e(φ) is interior to B and at e(φ) the vector
(g
t
0
, g
x
0
, g
t
1
, g
x
1
) (6.6 .11)
is neither zero nor orthogonal to B.
If we set
f
∗
(t) = A(