

Study with the several resources on Docsity
Earn points by helping other students or get them with a premium plan
Prepare for your exams
Study with the several resources on Docsity
Earn points to download
Earn points by helping other students or get them with a premium plan
The solution to homework set 6 in advanced optimal econometrics (aoe 5244), focusing on the extension required to include explicit dependence on time in the statement of an optimal control problem. An augmented state vector, transforms the problem to a standard form, and derives the variational hamiltonian and adjoint equations for the new problem. The document also explains how to find the transversality conditions for the extended costate.
Typology: Assignments
1 / 2
This page cannot be seen from the preview
Don't miss anything!


pendence on time (the independent variable) in the statement of the optimal
control problem. Specifically, we have dynamics given by
x˙ = f (x, t, u)
u ∈ Ω ⊂ IR
m , with given initial condition (x(t 0 ) = x 0 ), the end-condition
θ ı
(x(t f
), t f
) = 0, ı = 1, 2 ,... , q,
and the Mayer cost-functional
g(x(t f
), t f
To treat this problem we introduce an augmented state vector
xˆ =
[
x
x n+
]
and transform to a standard problem, where the t variable does not explicitly
appear. To this end we define
xˆ =
f (ˆx, u) =
[
f (x, x n+
, u)
]
and
θ ı
(x(t f
), t f
) = θ ı
(x(t f
), x n+
alongwith a modified Mayer cost-functional
gˆ(x(tf ), tf ).
The variational Hamiltonian for the new problem is written in terms of
the extended costate
λ = (λ,
λn+1)
λ, ˆx, u) = λ
T f (x, x n+
, u) +
λ n+
where the first term on the right is the variational Hamiltonian for the
original problem. The adjoint equations are
λ = −
(
∂ x
) T
(
∂ x
) T
λ n+
∂ x n+
∂ t
Since the modified problem does not explicitly include the independent vari-
abel (t), we have
H(t) = 0 constant.
Since
H is constant alongan extremal and since
λ n+
we find
d
d t
d H
d t
λ n+
or
d H
d t
λn+1 =
∂ t
The transversality conditions for the first n components of
λ are as be-
fore, but the last of these is
λ n+
(t f
) = λ 0
∂ ˆg
∂ x n+
q ∑
ı=
ν ı
θ ı
∂ x n+
Which can be written in terms of the original problem as
−H(t f
) = λ 0
∂ g
∂ t
q ∑
ı=
νı
∂ θ ı
∂ t