I NTRODUCTION
B ACKGROUND
M ETHODOLOGY
R ESULTS
C ONCLUSIONS
Optimal Control Model of Transmission of
Dengue: Theoretical Model
Dilruk Gallage & SSN Perera
University of Colombo
November 21, 2014
I NTRODUCTION
B ACKGROUND
M ETHODOLOGY
R ESULTS
O UTLINE P LAN
I NTRODUCTION
B ACKGROUND
M ETHODOLOGY
An Optimal Control Model of Dengue Disease
T HE F IRST- ORDER O PTIMALITY S YSTEM
R ESULTS
C ONCLUSIONS
C ONCLUSIONS
I NTRODUCTION
B ACKGROUND
M ETHODOLOGY
R ESULTS
C ONCLUSIONS
H OW THE OPTIMAL CONTROL THEORY IS USED IN
EPIDEMIOLOGY
Studies done by using optimal control
I
I
I
HIV diseases
Cancer
Hepatitis,...etc.
Limit the spread of diseases
I NTRODUCTION
B ACKGROUND
M ETHODOLOGY
R ESULTS
C ONCLUSIONS
A N O PTIMAL C ONTROL M ODEL OF D ENGUE D ISEASE
dSH
= (1 u1 )h SH bIV h SH
dt
dIH
= (1 u1 )h SH bIV (h + r)IH
dt
dRH
(1)
= rIH h RH
dt
dSV
= v (1 u2 )v SV bIH v SV
dt
dIV
= (1 u2 )v SV bIH v IV
dt
with initial conditions SH (0) = 0.9999, IH (0) = 0.0001, RH (0)=0,
SV (0)=1, and IV (0) = 0.
I NTRODUCTION
B ACKGROUND
M ETHODOLOGY
R ESULTS
C ONCLUSIONS
1 u1 (t): the efforts that prevent transmission of dengue virus
I
u1 = 1 the prevention of transmission of dengue virus is
100% effective,
u1 = 0 There is no prevention.
I NTRODUCTION
B ACKGROUND
M ETHODOLOGY
R ESULTS
C ONCLUSIONS
1 u1 (t): the efforts that prevent transmission of dengue virus
I
u1 = 1 the prevention of transmission of dengue virus is
100% effective,
u1 = 0 There is no prevention.
Consider the following cost functional
Z
Tf
1
ref
ref
(A1 (IH IH )2 +A2 (IV IV )2 + (B1 u21 +B2 u22 ))dt
2
0
(2)
subject to the state system (1).
J = J(y, u) =
A1 and A2 : the weight constants of infected humans and
infected mosquitoes respectively.
B1 and B2 : the weight on the cost of prevention and vector
control respectively.
I NTRODUCTION
B ACKGROUND
M ETHODOLOGY
R ESULTS
C ONCLUSIONS
T HE MAIN OBJECTIVE
minimizing the numbers of infected humans and infected
vectors
minimize J(y, u) with respect to u subject to (1)
Setting up the first order optimality by using Lagrangian
formulation.
(3)
I NTRODUCTION
B ACKGROUND
M ETHODOLOGY
R ESULTS
C ONCLUSIONS
S ETTING UP THE FIRST ORDER OPTIMALITY SYSTEM
We introduce the Lagrangian L : Y U Y R defined as
L(y, u, ) = J(y, u) + he(y, u), iY,Y ,
the first-order optimality system reads as
y,u, L(y, u, ) = 0.
We are able to obtain the following state system by considering
the variation of L with respect to the adjoint variable ,
dy
= f (y, u),
dt
(4)
with SH (0) = 0.9999, IH (0) = 0.0001, RH (0) = 0, SV (0) = 1, and
IV (0) = 0, where
I NTRODUCTION
B ACKGROUND
M ETHODOLOGY
R ESULTS
C ONCLUSIONS
(1 u1 )h SH bIV h SH
(1 u1 )h SH bIV (h + r)IH
.
rI
R
f (y, u) =
H
H
h
v (1 u2 )v SV bIH v SV
(1 u2 )v SV bIH v IV
Second, taking variations of L with respect to the state variable
y, we get the adjoint system,
d
= F(y, u, ),
dt
(5)
with SH (tf ) = 0, IH (tf ) = 0, RH (tf ) = 0, SV (tf ) = 0, IV (tf ) = 0,
where,
T
f
.
F(y, u, ) =
t
I NTRODUCTION
B ACKGROUND
M ETHODOLOGY
R ESULTS
C ONCLUSIONS
R ESULTS
0.45
Infected human proportion
0.4
Infected human proportion (IH) without controls
0.35
Infected human proportion (IH) with controls
0.3
0.25
0.2
0.15
0.1
0.05
0
50
100
150
200
250
Time (days)
Figure: IH with and without controls
300
350
I NTRODUCTION
B ACKGROUND
M ETHODOLOGY
R ESULTS
C ONCLUSIONS
Infected vector proportion
0.8
0.7
Infected vector proportion (I ) without controls
0.6
Infected vector proportion (IV) with controls
0.5
0.4
0.3
0.2
0.1
0
50
100
150
200
250
Time (days)
Figure: IV with and without controls
300
350
I NTRODUCTION
B ACKGROUND
M ETHODOLOGY
R ESULTS
C ONCLUSIONS
0.8
0.7
0.6
u1
0.5
0.4
0.3
0.2
0.1
0
50
100
150
200
Time (days)
Figure: The control u1
250
300
350
I NTRODUCTION
B ACKGROUND
M ETHODOLOGY
R ESULTS
C ONCLUSIONS
0.4
0.35
0.3
u2
0.25
0.2
0.15
0.1
0.05
0
50
100
150
200
Time (days)
Figure: The control u2
250
300
350
I NTRODUCTION
B ACKGROUND
M ETHODOLOGY
R ESULTS
C ONCLUSIONS
4
3.5
J(y,u)
3
2.5
2
1.5
1
0.5
10
Number of iterations
12
14
Figure: The cost functional J(y, u)
16
18
20
I NTRODUCTION
B ACKGROUND
M ETHODOLOGY
R ESULTS
C ONCLUSIONS
The SIR model extended by introducing two control
variables
A cost functional defined
The number of infected host and vector population is
reduced
C ONCLUSIONS
I NTRODUCTION
B ACKGROUND
M ETHODOLOGY
R ESULTS
C ONCLUSIONS
R EFERENCES
Perera, S.,S.,N., Controllability of Efficiency of Antiviral Therapy in Hepatitis
B Virus Infections, World Academy of Science, Engineering and
Technology, Vol.3, pp 938-942, 2009.
Gotz, T., Perera, S.,S.,N., Optimal Control of Melt Spinning Processes,
Journal of Engineering Mathematics, Vol.67, No.3, pp 153-163, 2010.
Jung, E., and Lenhart, S. et al., Optimal control of treatments in a two-strain
tuberculosis model, Discrete and continuous dynamical systems-series B,
Vol.2 pp. 473-482, 2002.
Pontryagin, L. S., Boltyanskii, V. G., Gamkrelidze, R. V., Mishchenko, E.
F. The Mathematical Theory of Optimal Processes, Wiley, New York, 1962.
Hattaf, K., Yous, N., Two optimal treatments of HIV infection model. World
Journal of Modelling and Simulation, Vol. 8 No. 1, pp. 27-35, 2012.
Pongsumpun, P., Transmission model for dengue disease with and without the
effect of extrinsic incubation period. KMITL Sci. Tech. J., Vol. 6 No. 2, pp
74-82, 2006.