‫الر ِح ِيم‬ ‫م‬‫َٰ‬

‫الر ْح ِ ه‬
‫ن‬ ‫َّللاِ ه‬
‫س ِم ه‬
‫ِب ْ‬
Proposal Defense:
‘’NUMERICAL ANALYSIS OF EPIDEMIC MODELS’’
Presented By
TAHIRA QAYYUM
MS-MATHEMATICS
Supervisor:
Dr. Shoaib Arif
Co-Supervisor:
Dr. Syed Ahmed Pasha
Department of Mathematics
Air University, Islamabad Pakistan

2
 Outline

INTRODUCTION
OBJECTIVES
LITERATURE REVIEW :
Selected paper 1
Selected paper 2
PERFORMANCE METRIC
METHODOLOGY
PRELIMINARY WORKk

3
 Introduction (1/3)

Mathematical epidemic model is a description of an infectious diseases.

It helps us to improve our understanding of transmission dynamics of
infectious diseases.
Two classes of epidemic models are deterministic epidemic models and the
stochastic epidemic models

4
 Introduction (2/3)
An epidemic model comprises compartments for example
Susceptible: The susceptible population who are not affected but
are at risk for infection.
Infectious: Infected and able to transmit the infection to other
susceptible individuals
Immune: Individuals who have either recovered and gained
permanent immunity from the disease SI S I
S – Susceptible
SIS S I S
I – Infectious
R - Removed
SIR S I R

5
 Introduction (3/3)
The assumptions are
• Population is fixed
• Infectious period = disease duration
• After recovery, individuals are immune
• The rate at which the total number of susceptible individuals become infected is
proportional to the product of susceptible and infected individuals

Diseases that can be modeled using the SIR are

flu
chicken-pox
rubella
small pox etc.

6
 Objectives

• Study of epidemic models for various diseases

• Simulation of spread of infectious diseases.
• Analysis of real data

7
 Literature Review(1/6)

1st Paper: 2nd Paper:

Nowzari, Cameron, Victor M.
Preciado, and George J. Pappas. Britton, Tom. "Stochastic
"Analysis and control of epidemics: A epidemic models: a survey."
survey of spreading processes on Mathematical biosciences 225.1
complex networks." IEEE Control (2010): 24-35.
Systems 36.1 (2016): 26-46.

8
 Literature Review(2/6)

SELECTED PAPER-1
Objective:
To study epidemic models on complex networks

9
 Literature Review(3/6)
SIR epidemic model
Assuming a large population size N
𝑁 = 𝑁 𝑆 + 𝑁𝐼 + 𝑁𝑅

𝑁𝑠 𝑆 𝑁
𝑖
𝐼 𝑁
𝑟
Define = 𝑁 , = 𝑁 , = 𝑁 𝑅 are fraction of susceptible, infected and recovered individuals
𝑁 𝑁 𝑁

So
𝑁 𝑆 + 𝑁𝐼 + 𝑁𝑅 = 1
dNS Β Infectious Rate
= −βNS NI (1)
dt γ Recovered Rate
dNI
= βNS NI − ϒNI (2) 𝑁𝑆 Susceptible Individuals
dt
dNR 𝑁𝐼 Infected Individuals
= ϒNI (3)
dt 𝑁𝑅 Recovered Individuals

10
From (1)
1 dNS
NI =− (4)
βNS dt

From (3)
I 1 dNR
N = (5)
ϒ dt

Equating (4) and (5)

1 𝑑N S 1 𝑑𝑁 𝑅
− =
𝛽N S 𝑑𝑡 ϒ 𝑑𝑡
1 dNS β dNR
=−
NS dt ϒ dt

11
Integrating both sides w.r.t ‘’t’’
β
ln N S − ln N0S = −
ϒ
β R
S − ϒ N
NS = N0 e
As N I = 1 − N R − N S
substituting in (3)
dt 1
=
dN R −
βNR
ϒ 1 − N R − N0S e ϒ

Integrating both sides w.r.t ‘’R’’.

1 1
t= ‫׬‬ βNR
d𝑅
ϒ −
1−NR −NS0 e ϒ

12
 Literature Review(4/6)
SIS epidemic model
𝑑NS
= ϒ𝑁 𝐼 + 𝛽N S 𝑁 𝐼 (6)
𝑑𝑡
𝑑𝑁𝐼
= 𝛽N S 𝑁 𝐼 − ϒ𝑁 𝐼 (7)
𝑑𝑡
Substituting N S = 1 − 𝑁 𝐼 in (7)
𝑑𝑁𝐼
− 𝛽 − ϒ 𝑁 𝐼 = −𝛽 𝑁 𝐼 2 (8)
𝑑𝑡
Bernoulli’s differential eq. is
𝑑𝑁𝐼
+ 𝑝 𝑡 𝑁 𝐼 = 𝑓 𝑡 (𝑁 𝐼 )𝑛
𝑑𝑡
Let 𝑁 𝐼 = 𝑢1−𝑛 = 𝑢 −1
𝑑𝑁𝐼 𝑑𝑁𝐼 𝑑𝑢
=
𝑑𝑡 𝑑𝑢 𝑑𝑡
Substituting in (8)
𝑑𝑢
−𝑢−2 − 𝛽 − ϒ 𝑢−1 = 𝛽𝑢−2
𝑑𝑡
𝑑𝑢
+ 𝛽−ϒ 𝑢 = 𝛽 (9)
𝑑𝑡
Which is linear differential eq.
So integrating factor= 𝑒 ‫𝛽 ׬‬−ϒ 𝑑𝑡 = 𝑒 𝛽−ϒ 𝑡

Eq.(9) becomes
𝑒 𝛽−ϒ 𝑡 𝑑𝑢 + 𝛽 − ϒ 𝑢𝑒 𝛽−ϒ 𝑡 = 𝛽𝑒 𝛽−ϒ 𝑡
𝑑𝑡
𝑑
𝑢𝑒 𝛽−ϒ 𝑡 = 𝛽𝑒 𝛽−ϒ 𝑡
𝑑𝑡
Integrating both sides
𝛽−ϒ 𝑡 𝑒 𝛽−ϒ 𝑡
𝑢𝑒 = 𝛽 + 𝐶1
𝛽−ϒ
𝛽 (𝑒 𝛽−ϒ 𝑡 +C)
𝑢=
𝛽−ϒ 𝑒 𝛽−ϒ 𝑡
1 1
𝐴𝑠 𝑁 𝐼 = ⇨𝑢=
𝑢 𝑁𝐼
ϒ 𝑒 𝛽−ϒ 𝑡
𝑁𝐼 𝑡 = 1−
𝛽 𝑒 𝛽−ϒ 𝑡 +C

14
 Literature Review(5/6)

SELECTED PAPER-2
Objectives:

• Study the dynamics of epidemic models

• Different analytical and numerical techniques are analyzed.
 Literature Review(6/6)

SIR stochastic epidemic model SIS epidemic model 𝑵𝑹 = 𝟎

𝑁 𝐼 𝑡 , 𝑁 𝑅 (𝑡) ∊ {0,1,2…….N}, and NI + 1 with probability
𝑁 𝑠 𝑡 = 𝑁 − 𝑁𝐼 𝑡 − 𝑁𝑅 𝑡 βNI N s ∆t + O ∆t
NI − 1 with probability
The state at some time t+∆t NI →
δNI ∆t + O ∆t
N I + 1, N R with probability NI with probability
βN I N s ∆t + O ∆t 1 − βN I N s + δN I ∆t + O ∆t
N I − 1, N R + 1 with probability
NI, NR →
δN I ∆t + O ∆t
NI, NR with probability
1 − βN I + δN I ∆t + O ∆t

16
Stochastic Epidemic Model (6/6)

SIR stochastic epidemic model

𝑁 𝐼 𝑡 , 𝑁 𝑅 (𝑡) ∊ {0,1,2…….N}
𝑁 𝑠 𝑡 = 𝑁 − 𝑁𝐼 𝑡 − 𝑁𝑅 𝑡

The state at some time t+∆t

17
 Stochastic epidemic model (6/6)

N I + 1, N R with probability
βN I N s ∆t + O ∆t
I R
N I − 1, N R + 1 with probability
N ,N →
δN I ∆t + O ∆t
NI , NR with probability
1 − βN I + δN I ∆t + O ∆t

For SIS Epidemic model N R = 0

NI + 1 with probability
βNI N s ∆t + O ∆t
NI − 1 with probability
NI →
δNI ∆t + O ∆t
NI with probability
1 − βN I N s + δN I ∆t + O ∆t

18
 Methodology

Numerical techniques for solving nonlinear equations for epidemic

model like Runge Kutta of order 4,Forward Euler,Non Standard Finite
Difference Scheme, Convergence Analysis of NSFD.

19
 Performance Metric

β
R0=
γ

R0 > 1 Infection spreads

R0 = 1 Infection remains constant
R0 < 1 Infection dies out

20
 Preliminary Work

Demonstrate the numerical techniques. Consider the SIR epidemic modelwe discretize
(1) and (2) by using
Forward Euler discretize model:
𝑁𝑆 𝑖+1 −𝑁𝑆 𝑖
From (1) = −𝛽𝑁 𝐼 𝑖 𝑁 𝑆 𝑖
∆𝑡

𝑁 𝑆 𝑖+1 = 𝑁 𝑆 𝑖 − ∆𝑡 𝛽𝑁 𝐼 𝑖 𝑁 𝑆 𝑖

𝑁𝐼 𝑖+1 −𝑁𝐼 𝑖
From (2) = 𝛽𝑁 𝐼 𝑖 𝑁 𝑆 𝑖 − ϒ𝑁 𝐼 𝑖
∆𝑡

𝑁 𝐼 𝑖+1 = 𝑁 𝐼 𝑖 + ∆𝑡 𝛽𝑁 𝐼 𝑖 𝑁 𝑆 𝑖 − ϒ𝑁 𝐼 𝑖

21
Initial conditions 𝐍𝟎𝐒 = 𝟒𝟎, 𝐍𝟎𝐈 = 𝟏𝟎

22
Initial conditions 𝐍𝟎𝐒 = 𝟒𝟎, 𝐍𝟎𝐈 = 𝟏𝟎

23
 Results

• When 𝑅0 ≻ 1 infectious individuals increase and susceptible

individuals decrease so disease will spread
• When 𝑅0 ≺ 1 infectious individuals decrease and susceptible
individuals increase so disease will not spread

24
Thank You

25