Appendices

155

7.A

UNIQUE OPTIMAL SOLUTION

According to the Pontryagin Maximum Principle [34], if u(·)is optimal for

the problem (7.1), (7.3) with the initial conditions given in Table 7.1 and fixed final

time tf, then there exists a nontrivial absolutely continuous mapping λ : [0,tf]R4,

λ(t) = (λ1(t)2(t)3(t)4(t)), called adjoint vector, such that

˙Sh = ∂H

∂λ1

,

˙Ih = ∂H

∂λ2

,

˙Sv = ∂H

∂λ3

,

˙Iv = ∂H

∂λ4

(7.A.1)

and

˙λ1 =∂H

∂Sh

,

˙λ2 =∂H

∂Ih

,

˙λ3 =∂H

∂Sv

,

˙λ4 =∂H

∂Iv

,

(7.A.2)

where function H defined by

H = H(Sh,Ih,Sv,Iv,λ,u)

= A1Ih + C

2 u2

+λ1h(1u)λhSh +γhIhµhSh)

+λ2 ((1u)λhSh(µh +γh +δh)Ih)

+λ3vλvSvµvbSv)

+λ4 (p2λvSvµvbIv)

is called the Hamiltonian, and the minimization condition

H(S

h(t),I

h(t),S

v(t),I

v(t)(t),u(t))

= min

0u1H(S

h(t),I

h(t),S

v(t),I

v(t)(t),u)

(7.A.3)

holds almost everywhere on [0,tf]. Moreover, the transversality conditions

λi(tf) = 0,

i = 1,...,4,

(7.A.4)

hold.

Theorem 7.A.1 Problem (7.1), (7.3) with fixed initial conditions Sh(0), Ih(0), Sv(0) and

Iv(0) and fixed final time tf, admits an unique optimal solution (S

h(·),I

h(·),S

v(·),I

v(·))

associated to an optimal control u(·) on [0,tf]. Moreover, there exists adjoint functions

λ

1(·), λ

2(·), λ

3(·) and λ

4(·) such that

˙λ

1(t) = λ

1(t)((1u(t))λh +µh)λ

2(t)λh(1u(t))

˙λ

2(t) =A1λ

1(t)γh +λ

2(t)(µh +γh +δh)

˙λ

3(t) = λ

3(t)(λv +µvb)λ

4(t)(λv))

˙λ

4(t) = λ

4(t)µvb ,

(7.A.5)