Optimal control of COVID-19 through strategic mathematical modeling: Incorporating harmonic mean incident rate and vaccination

Shah, Kamil; Shah, Jamal; Bonyah, Ebenezer; Alballa, Tmader; El-Wahed Khalifa, Hamiden Abd; Khan, Usman; Khan, Hameed

doi:10.1063/5.0229920

COVID-19 is a novel virus that has spread globally, and governments around the world often implement different strategies to prevent its spread. In the literature, several COVID-19 models have been studied with the bilinear incident rate. In this study, the S₁V₁E₁I₁Q₁R₁ (susceptible-vaccinated-exposed-infective-quarantined-recovered) COVID-19 model is proposed. To investigate how the disease spreads in the population, an algorithm is used. The efficacy of the algorithm is used to calculate the disease-free equilibrium point. A next generation matrix technique is used to find R₀. Furthermore, to check the effect of parameters on the basic reproduction number (R₀), the sensitivity analysis is conducted. Numerical simulation displays that the disease spreads in the population by increasing the value of the contact rate β while the disease spread in the population reduces by increasing the value of the vaccination rate θ, quarantine rate ϕ, and recovery rate γ. Different optimal control strategies, such as social distance and quick isolation, are also implemented.

I. INTRODUCTION

As of December 2019, COVID-19 had spread worldwide after first being reported in Wuhan, China. There have been many protective measures implemented against the transmission of this virus since it can be transmitted from person to person.¹ These precautions include the use of masks, maintaining physical distance, and the administration of vaccination in recent times. Symptoms of COVID-19 include high fever, dyspnea, and invasive multilobed lesions, as seen in chest radiographs.^2–9 According to reports, the symptoms of this virus appear for around five days after contracting COVID-19 infection.¹ The COVID-19 virus has been linked to hundreds of hospitalizations due to respiratory issues, flu-like symptoms, and other severe complications.¹ In some cases, these symptoms get worse with time, leading to death.²

In order to minimize COVID-19 infections, researchers are continuously working on developing effective vaccines and remedies. There have been a number of research articles with varying viewpoints published in the literature regarding minimizing COVID-19 infection. A mathematical model is the most effective tool for analyzing infectious disease transmission dynamics and evaluating the efficacy of interventions in reducing the prevalence. Many mathematical methods are under investigation for reducing the COVID-19 disease load.^10–14 Computational modeling was employed by Imai et al.^15,16 to estimate COVID-19’s prevalence in Wuhan based on the person-to-person transmission of the virus. Approximately 60% of the disease transmission can be prevented through preventive measures, according to their findings. According to Ngonghala et al.,¹⁷ COVID-19 can be reduced by maintaining social distance. Ndaïrou et al.¹⁸ proposed a COVID-19 model. In their study, they take super-spreaders spreading the disease among the population into account. Mandal et al.¹⁹ established a mathematical model to minimize the outbreak of COVID-19 by incorporating quarantine compartments and government intervention measures. Prathumwan et al.²⁰ investigated how confinement, hospitalization, and latent compartments affected COVID-19 transmission. Using a mask, social isolation, and early case detection, Okuonghae and Omame inspected the effectiveness of control measures in reducing COVID-19 dynamics.²¹ According to Khajanchi et al.,²² infection can be roughly divided into nine stages, and the most active method for stopping infection is the combination of pharmaceutical and non-pharmaceutical preventive measures. Rai et al.²³ used non-pharmaceutical interventions to study the usefulness of social media promotion to stop COVID-19 outbreaks. In order to define vaccination schedules, Acuña-Zegarra et al.²⁴ developed a mixed constraint optimal control problem. According to the findings, vaccines and reinfection periods are the main factors influencing COVID-19 mitigation. An investigation was conducted by Kurmi and Chouhan²⁵ using vaccination as a control to determine the outcome of immunization on COVID-19 outbreaks. Venkatesh and Ankamma Rao²⁶ investigated the impact of both effective and ineffective vaccination on the dynamics of COVID-19.

Numerous mathematical models have been examined based on the aforementioned literature in order to analyze the transmission dynamics of COVID-19 with the bilinear incidence rate. The transmission dynamics of COVID-19 with the effect of harmonic mean type incident for contact as well as for the vaccination rate have not been studied yet. As compared to other transmission rates, harmonic mean types are more realistic. In the present study, a harmonic mean type incident rate is considered for the transmission dynamics of COVID-19. In particular, the study uses harmonic rates to examine how vaccination affects COVID-19 transmission dynamics. Furthermore, different optimal control strategies, such as social distancing and quick isolation of infected individuals, are studied.

II. MODEL FORMULATION

An infectious disease spreading in a population is described by a system of differential equations. Each parameter in the equations represents a specific function in the context of the disease and represents the rate of change of different groups of individuals in the population. The total population N(t₁) = S₁(t₁) + V(t₁) + E₁(t₁) + I₁(t₁) + Q₁(t₁) + R₁(t₁) is subdivided into six classes: S₁(t₁) susceptible, V₁(t₁) vaccinated, E₁(t₁) exposed, I₁(t₁) infectious, Q₁(t₁) quarantined, and R₁(t₁) recovered population. Vaccinated and quarantined people are included in the model of COVID-19 presented in this study. The parameter Λ represents the birth or immigration rate of susceptible individuals into the population, while β is the transmission rate of the disease, indicating how likely an individual is to become infected through contact with an infected individual, The term $\frac{2 β S_{1} I_{1}}{S_{1} + I_{1}}$ represents the rate at which susceptible individuals become infected. Individuals susceptible to getting vaccinated are represented by λ, where μ is the natural death rate in each class. $\frac{2 β (1 - θ) V_{1} I_{1}}{V_{1} + I_{1}}$ represents the rate at which vaccinated individuals become infected. α₁ is the rate at which exposed individuals become infectious, while γ represents the infected people isolated from the society and become quarantined, and η is the rate at which infected people die due to the disease. Quarantined individuals get proper treatment during isolation, denoted by $γ (1 - δ) I_{1}$ ⁠, and transfer to the recovered class at a rate γδ. Infected individuals become healthy during quarantine while having proper treatment, denoted by ϕ, and transfer to the recovered class. The transmission dynamics of Eq. (1) are presented in Fig. 1.

FIG. 1.

A flow chart of S1V1E1I1Q1R1 COVID-19’s transmission dynamics.

View large Download slide

A flow chart of S₁V₁E₁I₁Q₁R₁ COVID-19’s transmission dynamics.

With the help of the above-mentioned flow chart, we construct the system of the differential S₁V₁E₁I₁Q₁R₁ model,

\begin{matrix} \frac{d S_{1}}{d t_{1}} = Λ - \frac{2 β S_{1} I_{1}}{S_{1} + I_{1}} - λ S_{1} - μ S_{1}, \\ \frac{d V_{1}}{d t_{1}} = λ S_{1} - \frac{2 β (1 - θ) V_{1} I_{1}}{V_{1} + I_{1}} - μ V_{1}, \\ \frac{d E_{1}}{d t_{1}} = \frac{2 β S_{1} I_{1}}{S_{1} + I_{1}} + \frac{2 β (1 - θ) V_{1} I_{1}}{V_{1} + I_{1}} - α_{1} E_{1} - μ E_{1}, \\ \frac{d I_{1}}{d t_{1}} = α_{1} E_{1} - γ I_{1} - η I_{1} - μ I_{1}, \\ \frac{d Q_{1}}{d t_{1}} = γ (1 - δ) I_{1} - ϕ Q_{1} - μ Q_{1}, \\ \frac{d R_{1}}{d t_{1}} = ϕ Q_{1} + γ δ I_{1} - μ R_{1} . \end{matrix}\}

(1)

Let us suppose that the initial conditions of system are of the form

S_{1} (0) \geq 0, V_{1} (0) \geq 0, E_{1} (0) \geq 0, I_{1} (0) \geq 0, Q_{1} (0) \geq 0, R_{1} (0) \geq 0 .

(2)

A. Positive characterization

Theorem 1.

The solution under consideration—positive $(S_{1}, E_{1}, V_{1}, I_{1}, Q_{1}, R_{o}) \in {R^{6}}_{+}$ of system (1)—is seen at any time t ≥ 0 when the rate of change of state variables is non-negative during the trivial stage.

Proof.

Observing system , the following are apparent:

\begin{matrix} {\frac{d S_{1}}{d t}|}_{S_{1} = 0} = Λ \geq 0, \\ {\frac{d V_{1}}{d t_{1}}|}_{V_{1} = 0} = λ S_{1} \geq 0, \\ {\frac{d E_{1}}{d t_{1}}|}_{E_{1} = 0} = \frac{2 β S_{1} I_{1}}{S_{1} + I_{1}} + \frac{2 β (1 - θ) V_{1} I_{1}}{V_{1} + I_{1}} \geq 0, \\ {\frac{d I_{1}}{d t_{1}}|}_{I_{1} = 0} = α_{1} E_{1} \geq 0, \\ {\frac{d Q_{1}}{d t_{1}}|}_{Q_{1} = 0} = γ (1 - δ) I_{1} \geq 0, \\ {\frac{d R_{1}}{d t_{1}}|}_{R_{1} = 0} = ϕ Q_{1} + γ δ I_{1} \geq 0 . \end{matrix}\}

(3)

1. Invariant region

The region in which the solution of model is uniformly bounded in the proper subset Π where Π ∈ R⁶ is

\begin{align} Π & = [(S_{1} (t_{1}), V_{1} (t_{1}), E_{1} (t_{1}), I_{1} (t_{1}), Q_{1} (t_{1}), \\ R_{1} (t_{1})) \in R_{+}^{6} : | N (t) \leq \frac{Λ}{μ}] . \end{align}

(4)

The entire population of the S₁V₁E₁I₁Q₁R₁ modal at time t₁ is specified as

\frac{d N (t_{1})}{d t_{1}} = \frac{S_{1} (t_{1})}{d t_{1}} + \frac{V_{1} (t_{1})}{d t_{1}} + \frac{E_{1} (t_{1})}{d t_{1}} + \frac{I_{1} (t_{1})}{d t_{1}} + \frac{Q_{1} (t_{1})}{d t_{1}} + \frac{R_{1} (t_{1})}{d t_{1}},

(5)

\begin{matrix} \frac{d N (t_{1})}{d t_{1}} = (Λ - \frac{2 β S_{1} I_{1}}{S_{1} + I_{1}} - λ S_{1} - μ S_{1}) \\ + (λ S_{1} - \frac{2 β (1 - θ) V_{1} I_{1}}{V_{1} + I_{1}} - μ V_{1}) \\ + (\frac{2 β S_{1} I_{1}}{S_{1} + I_{1}} + \frac{2 β (1 - θ) V_{1} I_{1}}{V_{1} + I_{1}} - α_{1} E_{1} - μ E_{1}) \\ + (α_{1} E_{1} - γ I_{1} - η I_{1} - μ I_{1}) \\ + (γ (1 - δ) I_{1} - ϕ Q_{1} - μ Q_{1}) \\ + (ϕ Q_{1} + γ δ I_{1} - μ R_{1}) \end{matrix}\},

(6)

\begin{matrix} \frac{d N (t_{1})}{d t_{1}} = Λ - \frac{2 β S_{1} I_{1}}{S_{1} + I_{1}} - λ S_{1} - μ S_{1} + λ S_{1} - \frac{2 β (1 - θ) V_{1} I_{1}}{V_{1} + I_{1}} - μ V_{1} \\ + \frac{2 β S_{1} I_{1}}{S_{1} + I_{1}} + \frac{2 β (1 - θ) V_{1} I_{1}}{V_{1} + I_{1}} - α_{1} E_{1} - μ E_{1} + α_{1} E_{1} - γ I_{1} \\ - η I_{1} - μ I_{1} + γ (1 - δ) I_{1} - ϕ Q_{1} - μ Q_{1} + ϕ Q_{1} + γ δ I_{1} - μ R_{1} \end{matrix}\} .

(7)

In addition, their total dynamics are given by

\frac{d N (t_{1})}{d t_{1}} = Λ - μ N (t_{1}) - η I_{1} (t_{1}),

(8)

\frac{d N (t_{1})}{d t_{1}} + μ N (t_{1}) \leq Λ,

(9)

that is,

\begin{matrix} 0 < N (S_{1} (t_{1}), V_{1} (t_{1}), E_{1} (t_{1}), I_{1} (t_{1}), Q_{1} (t_{1}), R_{1} (t_{1})) \leq \frac{Λ}{μ} (1 - e^{- μ t_{1}}) \\ + N (S_{1} (0), V_{1} (0), E_{1} (0), I_{1} (0), Q_{1} (0), R_{1} (0)) e^{- μ t_{1}} . \end{matrix}\} .

(10)

Now taking

\lim_{t_{1} \to \infty}

⁠, we get

0 \leq N \leq \frac{Λ}{μ}

⁠, which is positively invariant and bounded,

\begin{align} Π & = [(S_{1} (t_{1}), V_{1} (t_{1}), E_{1} (t_{1}), I_{1} (t_{1}), Q_{1} (t_{1}), \\ R_{1} (t_{1})) \in R_{+}^{6} : | N (t_{1}) \leq \frac{Λ}{μ}] . \end{align}

(11)

B. Disease-free equilibria and reproduction number

Utilizing model , we can show that it has a disease-free equilibrium, as indicated by E₀,

E^{0} = (S_{1}^{0}, V_{1}^{0}, E_{1}^{0}, I_{1}^{0}, Q_{1}^{0}, R_{1}^{0}) = (\frac{Λ}{λ + μ}, \frac{Λ}{μ (λ + μ)}, 0, 0, 0, 0) .

(12)

A disease reproduction number is crucial to understanding how diseases spread and how they can be controlled. When R₀ < 1, it indicates that the disease is dying out from the population. This helps in preventing the outbreak of epidemics. On the other hand, when R > 1, it indicates that the disease is spreading in the population. The next-generation matrix process

ρ (F * V^{- 1})

is used to compute the value of R₀ as follows:

X = (S_{1} (t_{1}), V_{1} (t_{1}), E_{1} (t_{1}), I_{1} (t_{1}), Q_{1} (t_{1}), R_{1} (t_{1})),

(13)

\frac{d X}{d t} = F - V,

(14)

F = [\begin{matrix} 0 \\ 0 \\ \frac{2 β S_{1} I_{1}}{S_{1} + I_{1}} + \frac{2 β (1 - θ) V_{1} I_{1}}{V_{1} + I_{1}} \\ \begin{matrix} 0 \\ 0 \\ 0 \end{matrix} \end{matrix}], V = [\begin{matrix} - Λ + \frac{2 β S_{1} I_{1}}{S_{1} + I_{1}} + λ S_{1} + μ S_{1} \\ - λ S_{1} + \frac{2 β (1 - θ) V_{1} I_{1}}{V_{1} + I_{1}} + μ V_{1} \\ α_{1} E_{1} + μ E_{1} \\ \begin{matrix} - α_{1} E_{1} + γ I_{1} + η I_{1} + μ I_{1} \\ - γ (1 - δ) I_{1} + ϕ Q_{1} + μ Q_{1} \\ - ϕ Q_{1} - γ δ I_{1} + μ R_{1} \end{matrix} \end{matrix}],

(15)

\begin{gathered} F = [\begin{matrix} 0 & 0 & 0 & 0 \\ 0 & 0 & 2 β + 2 β (1 - θ) & 0 \\ 0 & 0 & 0 & 0 \\ 0 & 0 & 0 & 0 \end{matrix}], \\ V = [\begin{matrix} μ & 0 & 2 β (1 - θ) & 0 \\ 0 & α_{1} + μ & 0 & 0 \\ 0 & - α_{1} & γ + η + μ & 0 \\ 0 & 0 & - γ (1 - δ) & ϕ + μ \end{matrix}] . \end{gathered}

(16)

In addition,

V^{- 1} = [\begin{matrix} \frac{γ + η + μ}{γ μ + η μ + μ^{2}} & - \frac{2 α_{1} β [1 - θ]}{(γ μ + η μ + μ^{2}) (μ + α_{1})} & - \frac{2 β [1 - θ]}{γ μ + η μ + μ^{2}} & 0 \\ 0 & \frac{1}{μ + α_{1}} & 0 & 0 \\ 0 & \frac{μ α_{1}}{(γ μ + η μ + μ^{2}) (μ + α_{1})} & \frac{μ^{2} + μ α_{1}}{(γ μ + η μ + μ^{2}) (μ + α_{1})} & 0 \\ 0 & \frac{μ α_{1} γ [1 - δ]}{(γ μ + η μ + μ^{2}) (μ + ϕ) (μ + α_{1})} & \frac{μ^{2} γ [1 - δ] + μ α_{1} γ [1 - δ]}{(γ μ + η μ + μ^{2}) (μ + ϕ) (μ + α_{1})} & \frac{1}{μ + ϕ} \end{matrix}],

(17)

\begin{align} F V^{- 1} & = [\begin{matrix} 0 & 0 & 0 & 0 \\ 0 & 0 & 2 β + 2 β (1 - θ) & 0 \\ 0 & 0 & 0 & 0 \\ 0 & 0 & 0 & 0 \end{matrix}] [\begin{matrix} \frac{γ + η + μ}{γ μ + η μ + μ^{2}} & - \frac{2 α_{1} β [1 - θ]}{(γ μ + η μ + μ^{2}) (μ + α_{1})} & - \frac{2 β [1 - θ]}{γ μ + η μ + μ^{2}} & 0 \\ 0 & \frac{1}{μ + α_{1}} & 0 & 0 \\ 0 & \frac{μ α_{1}}{(γ μ + η μ + μ^{2}) (μ + α_{1})} & \frac{μ^{2} + μ α_{1}}{(γ μ + η μ + μ^{2}) (μ + α_{1})} & 0 \\ 0 & \frac{μ α_{1} γ [1 - δ]}{(γ μ + η μ + μ^{2}) (μ + ϕ) (μ + α_{1})} & \frac{μ^{2} γ [1 - δ] + μ α_{1} γ [1 - δ]}{(γ μ + η μ + μ^{2}) (μ + ϕ) (μ + α_{1})} & \frac{1}{μ + ϕ} \end{matrix}] . \end{align}

(18)

After simplification, the basic reproduction is

R_{0} = \frac{2 β α_{1} (2 - θ)}{(μ + α_{1}) (η + μ + γ)} .

(19)

III. LOCAL STABILITY OF DISEASE-FREE EQUILIBRIA

Theorem 3.1.

When R₀ < 1, the SIR model (1) is stable locally asymptotically at disease-free equilibrium E⁰.

Proof.

When the left-hand side of the Jacobian matrix at J(E₀) is set to zero, we obtain the Jacobian matrix of model , which gives

J (E^{0}) = (\begin{matrix} - λ - μ & 0 & 0 & - 2 β & 0 & 0 \\ λ & - μ & 0 & - 2 β (1 - θ) & 0 & 0 \\ 0 & 0 & - μ - α_{1} & 2 β + 2 β (1 - θ) & 0 & 0 \\ 0 & 0 & α_{1} & - γ - η - μ & 0 & 0 \\ 0 & 0 & 0 & γ (1 - δ) & - μ - ϕ & 0 \\ 0 & 0 & 0 & γ δ & ϕ & - μ \end{matrix}),

(20)

|J (E^{0}) - X_{1}| = 0,

(21)

|\begin{matrix} a_{1} - ℘_{o} & 0 & 0 & - 2 β & 0 & 0 \\ λ & a_{2} - ℘_{o} & 0 & - 2 β (1 - θ) & 0 & 0 \\ 0 & 0 & a_{3} - ℘_{o} & 2 β + 2 β (1 - θ) & 0 & 0 \\ 0 & 0 & α_{1} & a_{4} - ℘_{o} & 0 & 0 \\ 0 & 0 & 0 & γ (1 - δ) & a_{5} - ℘_{o} & 0 \\ 0 & 0 & 0 & γ δ & ϕ & a_{2} - ℘_{o} \end{matrix}| = 0,

(22)

where

\begin{matrix} a_{1} = - (λ + μ), \\ a_{2} = - μ, \\ a_{3} = - (μ + α_{1}), \\ a_{4} = - (γ + η + μ), \\ a_{5} = - (μ + ϕ), \end{matrix}\}

\begin{align} (a_{2} - ℘_{o}) (a_{1} - ℘_{o}) (a_{2} - ℘_{o}) (a_{5} - ℘_{o}) \\ \times |\begin{matrix} a_{3} - ℘_{o} & 2 β + 2 β (1 - θ) \\ α_{1} & a_{4} - ℘_{o} \end{matrix}| = 0, \end{align}

(23)

\begin{align} (a_{2} - ℘_{o}) (a_{1} - ℘_{o}) (a_{2} - ℘_{o}) (a_{5} - ℘_{o}) ((a_{3} - ℘_{o}) (a_{4} - ℘_{o}) \\ - α_{1} (2 β + 2 β (1 - θ))) = 0, \end{align}

(24)

\begin{align} (a_{2} - ℘_{o}) (a_{1} - ℘_{o}) (a_{2} - ℘_{o}) (a_{5} - ℘_{o}) (a_{3} a_{4} - a_{3} ℘_{o} - a_{4} ℘_{o} + {℘_{o}}^{2} \\ - α_{1} (2 β + 2 β (1 - θ))) = 0, \end{align}

(25)

\begin{align} (a_{2} - ℘_{o}) (a_{1} - ℘_{o}) (a_{2} - ℘_{o}) (a_{5} - ℘_{o}) ({℘_{o}}^{2} - (a_{3} + a_{4}) ℘_{o} \\ + a_{3} a_{4} - α_{1} (2 β + 2 β (1 - θ))) = 0 . \end{align}

(26)

Clearly,

℘_{o} = - μ, ℘_{o} = - (λ + μ), ℘_{o} = - μ, ℘_{o} = - (μ + ϕ)

are negative.

According to the Routh–Hurwitz criterion of order two, $(γ + η + 2 μ + α_{1}) > 0$ and $- 2 β α_{1} (2 - θ) + (μ + α_{1}) (γ + η + μ) > 0$ ⁠.

From

- 2 β α_{1} (2 - θ) + (μ + α_{1}) (γ + η + μ) > 0

⁠,

\begin{matrix} - 2 β α_{1} (2 - θ) + (μ + α_{1}) (γ + η + μ) > 0, \\ (μ + α_{1}) (γ + η + μ) > 2 β α_{1} (2 - θ), \\ 1 > \frac{2 β α_{1} (2 - θ)}{(μ + α_{1}) (γ + η + μ)}, \\ 1 > R_{0} . \end{matrix}\}

(27)

R_o < 1, so DFE is locally asymptotically stable.

IV. SENSITIVITY ANALYSIS

A model of epidemics must include R₀. If the R₀ value is huge, then it is challenging to control a disease. By targeting the parameters that produce a greater deviation in the basic reproduction number, we aim to reach a greater level of reproducibility. Variations in a parameter can cause the associated variance in a state variable. Based on the definition given in Ref. 27, these indices have been calculated. Given below is the partial derivative representation of the sensitivity index:

ϒ_{p}^{R_{0}} = \frac{p}{R_{0}} \times \frac{\partial R_{0}}{\partial p},

(28)

where p denotes the parameters and Ro denotes the reproductive number.

The sensitivity analysis shown in Table I demonstrates that certain parameters are more sensitive to reproduction numbers than others. μ, θ, α₁, η, and γ reduce the spread of infection in the population, while on the other hand, β increases the spread of the infection when it gets high (Fig. 2).

TABLE I.

Sensitivity indices of R₀.

No.	Parameters	Analytical values	Numerical values
1	β	1	1.000 00
2	μ	$- \frac{μ}{γ + η + μ} - \frac{μ}{μ + α_{1}}$	−0.947 804 993 048 809 4
3	θ	$\frac{θ}{- 2 + θ}$	−0.052 631 578 947 368 42
4	α₁	$\frac{μ}{μ + α_{1}}$	0.460 000 115 000 028 8
5	γ	$- \frac{γ}{γ + η + μ}$	−0.365 853 658 536 585 4
6	η	$- \frac{η}{γ + η + μ}$	−0.146 341 463 414 634 17

No.	Parameters	Analytical values	Numerical values
1	β	1	1.000 00
2	μ	$- \frac{μ}{γ + η + μ} - \frac{μ}{μ + α_{1}}$	−0.947 804 993 048 809 4
3	θ	$\frac{θ}{- 2 + θ}$	−0.052 631 578 947 368 42
4	α₁	$\frac{μ}{μ + α_{1}}$	0.460 000 115 000 028 8
5	γ	$- \frac{γ}{γ + η + μ}$	−0.365 853 658 536 585 4
6	η	$- \frac{η}{γ + η + μ}$	−0.146 341 463 414 634 17

FIG. 2.

View large Download slide

Sensitivity indices of R₀.

V. NUMERICAL SIMULATION OF SVEIQR MODEL

The following figures show the effects of primary and secondary parameters of the S₁V₁E₁I₁Q₁R₁ model specially on the infectious class to discuss and observe the effects of different parameters. In addition, we find the range of the following parameters to find if the effects are directly or inversely proportional to the infectious class and if the disease spreads to or emanates from the population.

Figure 3(a) depicts the relationship with the contact rate β. As the contact rate beta increases, the susceptible population likely decreases, indicating a faster spread of disease due to more frequent contacts. In Fig. 3(b), a higher β shows an increase in exposed individuals, representing an increase in people who have contracted the disease but are not yet infectious. Both graphs illustrate the impact of the contact rate on disease dynamics within a population.

FIG. 3.

View large Download slide

(a) The influence of β on susceptible individuals. (b) The influence of β on exposed individuals.

As shown in Fig. 4(a), as the vaccination rate λ increases, the proportion of the susceptible population is expected to decrease. This decline occurs because more individuals are receiving vaccinations, reducing the number of people vulnerable to the disease. In contrast, Fig. 4(b) displays a direct positive correlation: as λ increases, the number of vaccinated individuals in the population increases accordingly. These graphs together illustrate the critical role of vaccination in controlling disease spread by decreasing susceptibility and increasing immunity within a population.

FIG. 4.

View large Download slide

(a) The impact of λ on susceptible individuals. (b) The impact of λ on vaccinated individuals.

Figure 5(a) shows that as θ increases, the vaccine effectiveness decreases, potentially leading to more breakthrough infections, but the total number of vaccinated individuals remains unaffected by θ. Meanwhile, Fig. 5(b) shows an upsurge in the number of exposed individuals within the vaccinated population as θ increases. This indicates that higher rates of infection among vaccinated individuals correspond to more of them being exposed to the disease. These graphs collectively highlight the importance of vaccine efficacy and the implications of breakthrough infections in understanding disease dynamics in vaccinated populations.

FIG. 5.

View large Download slide

(a) The influence of θ on exposed individuals. (b) The influence of θ on vaccinated individuals.

Figure 6(a) illustrates how changes in α₁ affect exposed individuals. An increase in α₁ might indicate a faster transition from being exposed to becoming infectious, which could initially lead to a decrease in the exposed population as they quickly move to the infected category. Figure 6(b) shows that a higher α₁ would typically correspond to an increase in the infected population. This is because a greater α₁ means exposed individuals are becoming infectious more rapidly, thereby swelling the ranks of the infected group. Together, these graphs demonstrate the dynamics of disease progression within a population, particularly how changes in the rate of progression from exposed to infectious impact both the exposed and infected populations. It underscores the importance of this transition rate, considering and managing the spread of infectious diseases.

FIG. 6.

(a) The impact of α1 on exposed individuals. (b) The impact of α1 on infected individuals.

View large Download slide

(a) The impact of α₁ on exposed individuals. (b) The impact of α₁ on infected individuals.

Figure 7(a) exhibits the impact of varying recovery rates on the infected population over time. As the recovery rate γ increases from 0.05 to 0.21, the peak of the infected population occurs earlier and at a lower level, indicating that a higher recovery rate can significantly reduce the spread of an infection and shorten the duration of an epidemic. Figure 7(b) depicts the trajectory of a quarantined population over time, given different values of parameter γ. Each curve represents a scenario with a distinct γ value, showing how the size of the quarantined population changes over a period of 100 days. The peak of each curve suggests the maximum number of people quarantined at any given time, and this peak occurs at different times depending on the value of γ. Higher values of γ result in a quicker and lower peak, indicating a faster rate of quarantine and a smaller maximum quarantined population. Conversely, lower γ values lead to a slower response and a higher peak. After reaching the maximum, each curve shows a decline, suggesting that fewer people remain in quarantine over time.

FIG. 7.

View large Download slide

(a) The impact of γ on infected individuals. (b) The impact of γ on quarantine individuals.

Figure 8(a) exhibits an increase in ϕ (the rate of recovery of an infected individual becoming healthy during quarantine), which might initially lead to a larger quarantine population due to effective detection and isolation. However, as ϕ becomes higher, indicating more successful treatments, the duration of individuals in quarantine could shorten, affecting the overall quarantine population count. Conversely, Fig. 8(b) shows a clear positive trend, with higher ϕ values leading to a greater number of recoveries. This indicates that effective treatment and speedy recovery from infection significantly increase the recovered population, highlighting the crucial role of medical intervention in managing disease outbreaks.

FIG. 8.

View large Download slide

(a) The impact of ϕ on infected individuals. (b) The impact of ϕ on quarantine individuals.

Figure 9(a) shows that as δ (effectiveness of treatment for quarantined individuals) increases, the quarantine population might initially increase due to more effective isolation. However, with better treatment, the duration in quarantine may decrease, leading to quicker recoveries and a fluctuating quarantine population. Figure 9(b) shows a positive trend, with an increase in δ resulting in more individuals recovering from the disease. This suggests that efficient and effective treatment significantly boosts the recovery rate, highlighting the vital role of medical care in managing disease outbreaks and improving patient outcomes.

FIG. 9.

View large Download slide

(a) The impact of δ on infected individuals. (b) The impact of δ on quarantine individuals.

VI. FORMULATION OF THE OPTIMAL CONTROL SVEIQR MODEL

We can design a control strategy for overcoming this pandemic built on the sensitivity analysis discussed in the previous section. The estimated parameters were analyzed using sensitivity analysis when we compared actual data with the model output. The future behavior of this pandemic can thus be predicted better. Using model , we reshape it by observing the most parameters to investigate the influence of control measures on the future scenario. In model , two controls are included, and the result is as follows:

\begin{matrix} \frac{d S_{1}}{d t_{1}} = Λ - \frac{2 β (1 - u_{1}) S_{1} I_{1}}{S_{1} + I_{1}} - λ S_{1} - μ S_{1}, \\ \frac{d V_{1}}{d t_{1}} = λ S_{1} - \frac{2 β (1 - θ) V_{1} I_{1}}{V + I} - μ V_{1}, \\ \frac{d E_{1}}{d t_{1}} = \frac{2 β (1 - u_{1}) S_{1} I_{1}}{S_{1} + I_{1}} + \frac{2 β (1 - θ) V_{1} I_{1}}{V_{1} + I_{1}} - α_{1} E_{1} - μ u_{2} E_{1} - ω u_{2} E_{1}, \\ \frac{d I_{1}}{d t_{1}} = α_{1} E_{1} - γ I_{1} - η I_{1} - μ I_{1}, \\ \frac{d Q_{1}}{d t_{1}} = γ (1 - δ) I_{1} - ϕ Q_{1} - μ Q_{1} + ω u_{2} E_{1}, \\ \frac{d R_{1}}{d t_{1}} = ϕ Q_{1} + γ δ I_{1} - μ R_{1} . \end{matrix}\}

(29)

Preventative measures represented by control u₁(t) include social distancing and the frequent use of masks. The aim is to prevent infection from spreading to healthy individuals. COVID-19 virus isolates people quickly by using control u₂(t). A objective functional J is developed to determine the best methods for controlling the disease. In addition to reducing infectious individuals, u₁(t) and u₂(t) are focused on minimizing the cost of implementation,

J (u_{1}, u_{2}) = \int_{0}^{T_{f}} [M_{1} E_{1} + M_{2} I_{1} + M_{2} Q_{1} + C_{1} u_{1}^{2} + C_{2} u_{2}^{2}] d t,

(30)

where M₁, M₂, and M₃ describe the positive weights. The objective function we have described aims to reduce the costs of the above-mentioned controls while reducing the number of infections. There are two controls,

u_{1}^{*}

and

u_{2}^{*}

⁠, which can be found as follows:

J (u_{1}^{*}, u_{2}^{*}) = \min [J (u_{1}^{2}, u_{2}^{2}), (u_{1}^{2}, u_{2}^{2}) \in U],

(31)

according to which we have U, which is called the control set and is described by

U = [(u_{1}^{2}, u_{2}^{2}) : i s m e a s u r a b l e a n d [0, 1], 0 \leq u_{i} (t) \leq 1 for i = 1, 2] .

(32)

Our essential conditions and solution to can be obtained by using the Pontryagin’s maximum principle.²⁸

A. An analysis of optimal control and its existence

With the aid of powerful classical procedures, it is possible to verify the existence and analysis of an optimal control. The following hypotheses should be examined according to Ref. 29:

H1: The control set and corresponding variables described in the hypothesis are nonempty.
H2: The convexity and closeness properties of U must be satisfied.
H₃: The described system is bounded in the state of control by a linear function on the right-hand side.
H₄: A convex integrand on U bounds the objective functional J,
$d_{1} {(\sum_{i = 1}^{3} {|u_{i}|}^{2})}^{2} - d_{2}, w h e r e d_{1}, d_{2} > 0, a n d t_{1} > 1 .$
(33)
According to Lukes,³⁰ the solution in system (29) exists if the result is valid. Thus, the hypothesis given above can be verified. Bounded coefficients and solutions are in accordance with H1, and boundedness demonstrates that U fulfills the second hypothesis. The right-hand side of (29) fulfills the third hypothesis H3 because system (39) is bilinear in u₁ and u₂ and because the solutions are bounded. We can choose constants M₁, M₂, M₃, C₁, C₂, d₁, and d₂ that are all positive and t₁ > 1. The inequality holds since control functions u₁ and u₂ are convex and the number of infectious hosts and vectors is also convex,
$M_{1} E_{1} + M_{2} I_{1} + M_{3} Q_{1} + C_{1} u_{1}^{2} + C_{2} u_{2}^{2} \geq d_{1} {(\sum_{i = 1}^{3} {|u_{i}|}^{2})}^{\frac{t_{1}}{2}} - d_{2} .$
(34)
Thus, we can also verify the last condition. As a result, we arrive at the following theorem:
$\begin{align} L (E_{1}, I,_{1} Q_{1}, u_{1}, u_{2}) & = M_{1} E_{1} + M_{2} I_{1} + M_{3} Q_{1} + C_{1} u_{1}^{2} \\ + C_{2} u_{2}^{2} \geq d_{1} ({|u_{1}|}^{2} + {|u_{2} - 2|}^{2}) . \end{align}$
(35)
Based on the above-mentioned formula, one must construct the least possible value. Defining the Hamiltonian this way gets us to objective results: Let us examine $X = (S_{1}, V_{1}, E_{1}, I_{1}, Q_{1}, R_{1}), U = (u_{1}, u_{2})$ and $ψ = (ψ_{1} (t_{1}),$ $ψ_{2} (t_{1}), ψ_{3} (t_{1}), ψ_{4} (t_{1}), ψ_{5} (t_{1}), ψ_{6} (t_{1}))$ ⁠, with the intention of obtaining
$\begin{matrix} H (X, U, ψ) = M_{1} E_{1} + M_{2} I_{1} + M_{3} Q_{1} + C_{1} u_{1}^{2} + C_{2} u_{2}^{2} \\ + ψ_{1} (t_{1}) [Λ - \frac{2 β (1 - u_{1}) S_{1} I_{1}}{S_{1} + I_{1}} - λ S_{1} - μ S_{1}] \\ + ψ_{2} (t_{1}) [λ S_{1} - \frac{2 β (1 - θ) V_{1} I_{1}}{V_{1} + I_{1}} - μ V_{1}] \\ + ψ_{3} (t_{1}) [\begin{matrix} \frac{2 β (1 - u_{1}) S_{1} I_{1}}{S_{1} + I_{1}} + \frac{2 β (1 - θ) V_{1} I_{1}}{V_{1} + I_{1}} \\ - α_{1} E_{1} - μ u_{2} E_{1} - ω u_{2} E_{1} \end{matrix}] \\ + ψ_{4} (t_{1}) [α_{1} E_{1} - γ I_{1} - η I_{1} - μ I_{1}] \\ + ψ_{5} (t_{1}) [γ (1 - δ) I_{1} - ϕ Q_{1} - μ Q_{1} + ω u_{2} E_{1}] \\ + ψ_{6} (t_{1}) [ϕ Q_{1} + γ δ I_{1} - μ R_{1}], \end{matrix}\} .$
(36)
For determining the essential optimal control conditions, Pontryagin’s maximum principle is used as reported in Ref. 31. Using the following formula, this can be calculated: For optimized solution $(u_{1}^{*}, u_{2}^{*})$ of optimal control problem (30), a vector function $ψ = (ψ_{1} (t_{1}), ψ_{2} (t_{1}),$ $ψ_{3} (t_{1}), ψ_{4} (t_{1}), ψ_{5} (t_{1}), ψ_{6} (t_{1}))$ ⁠, which is nontrivial, prevails, which fulfills the ensuing conditions. In the statement, it is stated that
$\begin{gathered} \frac{\partial x}{d t_{1}} = \frac{\partial}{\partial ψ} (H (t_{1}, U^{*}, ψ (t_{1}))), 0 = \frac{\partial}{\partial u} (H (t_{1}, U^{*}, ψ (t_{1}))), \\ \frac{\partial ψ}{d t_{1}} = - \frac{\partial}{\partial x} (H (t_{1}, U^{*}, ψ (t_{1}))) . \end{gathered}$
(37)
Currently, H is subject to mandatory conditions.

Theorem 7.1.1.

For $U = (u_{1}^{*}, u_{2}^{*})$ and solution $X = (S_{1}, V_{1},$ $E_{1}, I_{1}, Q_{1}, R_{1})$ of system (1), variables $ψ = (ψ_{1} (t_{1}), ψ_{2} (t_{1}), ψ_{3} (t_{1}),$ $ψ_{4} (t_{1}), ψ_{5} (t_{1}), ψ_{6} (t_{1}))$ ⁠.

Proof.

The adjoint equations and transversally related conditions can be obtained by utilization. Set

S_{1} = S_{1}^{*}, V_{1} = V_{1}^{*}, E_{1} = E_{1}^{*}, I_{1} = I_{1}^{*}, Q_{1} = Q_{1}^{*}, R_{1} = R_{1}^{*}

and differentiate the Hamiltonian H with describe to variables

X = (S_{1}, V_{1}, E_{1}, I_{1}, Q_{1}, R_{1})

⁠, Hence, we have

\{\begin{cases} \frac{d ψ_{1}}{d t_{1}} = \frac{(2 I_{1}^{2} β u_{1} - 2 I_{1} S_{1} (λ + μ) - S_{1}^{2} (λ + μ) - I_{1}^{2} (2 β + λ + μ)) ψ_{1} + {(I_{1} + S_{1})}^{2} λ ψ_{2} - 2 I_{1}^{2} β (u_{1} - 1) ψ_{3}}{{(I_{1} + S_{1})}^{2}}, \\ \frac{d ψ_{2}}{d t_{1}} = \frac{(2 I_{1}^{2} β (θ - 1) - {(I + V_{1})}^{2} μ) ψ_{2} - 2 I_{1}^{2} β (θ - 1) ψ_{3}}{{(I_{1} + V_{1})}^{2}}, \\ \frac{d ψ_{3}}{d t_{1}} = M_{1} - (μ + w u_{2} + α_{1}) ψ_{3} + α_{1} ψ_{4} + w u_{2} ψ_{5}, \\ \frac{d ψ_{4}}{d t_{1}} = γ δ ψ_{6} - γ (δ - 1) ψ_{5} - (γ + η + μ) ψ_{4} + \frac{2 S_{1}^{2} β (u_{1} - 1) ψ_{2}}{{(I + S_{1})}^{2}} \\ + 2 β ψ_{3} (\frac{S_{1} I_{1} (u_{1} - 1)}{{(I_{1} + S_{1})}^{2}} + \frac{S_{1} (1 - u_{1})}{I_{1} + S_{1}} + \frac{(θ - 1) I_{1} V_{1}}{{(I_{1} + V_{1})}^{2}} + \frac{(1 - θ) V_{1}}{I_{1} + V_{1}}) + \frac{2 β (θ - 1) V_{1}^{2} ψ_{2}}{{(I_{1} + V_{1})}^{2}} + M_{2}, \\ \frac{d ψ_{5}}{d t_{1}} = M_{3} + (- μ - ϕ) ψ_{5} + ϕ ψ_{6}, \\ \frac{d ψ_{6}}{d t_{1}} = - μ ψ_{6}, \end{cases}

(38)

with transverse conditions ψ₁(T) = ψ₂(T) = ψ₃(T) = ψ₄(T) = ψ₅(T) = ψ₆(T) = 0. As a result, the property set U of controls and the optimality conditions give the result

u_{1}^{*} = \max \{\min \{1, \frac{(ψ_{1} - ψ_{3}) β_{1} S_{1} I_{1}}{C_{1} (I_{1} + S_{1})}, 0\}\},

(39)

u_{2}^{*} = \max \{\min \{1, \frac{ω E_{1} (ψ_{3} - ψ_{5})}{2 C_{2}}, 0\}\} .

(40)

There is also a numerical method for observing the outcome of optimal controls on COVID-19 transmission.

B. Numerical simulation of optimal control

Parameters used for simulated optimal control in this subsection are mentioned in Table II. Moreover, we compare the results of the optimal control applied to the time intervals without and with control. Indicated by the weight of the treatment. control u₁ and weight accompanying the infected population u₂ are taken as unity. In this second attempt at solving the optimal system (10), we will use iterative Runge–Kutta procedures of fourth order. A Runge–Kutta fourth order procedure is applied to solve state (10), based on our initial guess for the state variables. Our primary objective is to address the deputy variables through backward Runge–Kutta fourth-order procedures while simultaneously including state variables and transverse conditions. In addition to providing a comprehensive understanding of system dynamics, this approach enhances the accuracy of our analysis.

TABLE II.

Assumed parameter values of the SVEIQR COVID-19 disease model.

Parameters	Estimated values	References
Λ	0.464	32
β	0.503	33
λ	0.4	32
μ	0.04	Assumed
θ	0.1	Assumed
α₁	0.046 956 5	Assumed
γ	0.030	Assumed
η	0.012	33
δ	0.2	Assumed
ϕ	0.082	34
ω	0.3	Assumed

Parameters	Estimated values	References
Λ	0.464	32
β	0.503	33
λ	0.4	32
μ	0.04	Assumed
θ	0.1	Assumed
α₁	0.046 956 5	Assumed
γ	0.030	Assumed
η	0.012	33
δ	0.2	Assumed
ϕ	0.082	34
ω	0.3	Assumed

Figure 10(a) shows the number of people who are susceptible. Without interventions, the number of susceptible people decreases more slowly over time (red curve), indicating a slower disease spread. With interventions (blue curve), the number of susceptible people decreases more rapidly, suggesting that the interventions are operative in dropping the disease outbreak.

FIG. 10.

(a) Comparison of two scenarios in a disease outbreak SVEIQR model over time, one without interventions (red curve, u1 = 0, u2 = 0) and one with interventions (blue curve, u1 ≠ 0, u2 ≠ 0). (b) Comparison of two scenarios in a disease outbreak SVEIQR model over time, one without interventions (red curve, u1 = 0, u2 = 0) and one with interventions (blue curve, u1 ≠ 0, u2 ≠ 0).

View large Download slide

(a) Comparison of two scenarios in a disease outbreak SVEIQR model over time, one without interventions (red curve, u₁ = 0, u₂ = 0) and one with interventions (blue curve, u₁ ≠ 0, u₂ ≠ 0). (b) Comparison of two scenarios in a disease outbreak SVEIQR model over time, one without interventions (red curve, u₁ = 0, u₂ = 0) and one with interventions (blue curve, u₁ ≠ 0, u₂ ≠ 0).

Figure 10(b) shows the number of vaccinated individuals. Without optimal control, the number of vaccinated people remains at zero (red curve), as expected. With optimal control, the number of vaccinated people increases and then decreases over time (blue curve), reflecting the implementation of a vaccination campaign. After reaching a peak, it declines as fewer susceptible people are left to vaccinate.

Figure 11(a) illustrates the population that has encountered the infection but has not yet reached the stage of being contagious. The red curve demonstrates a rapid increase in exposed individuals when no control measures are in place, reaching a peak before declining. The blue curve, where control measures are u₁ ≠ 0 and u₂ ≠ 0, shows a slower increase in exposed individuals, indicating that the measures are delaying the onset of the disease and reducing the number of exposed individuals.

FIG. 11.

(a) The dynamics of an epidemic over time, comparing two scenarios: without control measures (red curve u1 = 0, u2 = 0) and with control measures (blue curve u1 ≠ 0, u2 ≠ 0). (b) The dynamics of an epidemic over time, comparing two scenarios: without control measures (red curve u1 = 0, u2 = 0) and with control measures (blue curve u1 ≠ 0, u2 ≠ 0).

View large Download slide

(a) The dynamics of an epidemic over time, comparing two scenarios: without control measures (red curve u₁ = 0, u₂ = 0) and with control measures (blue curve u₁ ≠ 0, u₂ ≠ 0). (b) The dynamics of an epidemic over time, comparing two scenarios: without control measures (red curve u₁ = 0, u₂ = 0) and with control measures (blue curve u₁ ≠ 0, u₂ ≠ 0).

Figure 11(b)’s graph represents the number of people who are infectious and can spread the disease. Similar to the left graph, the red curve without control measures shows a sharp increase to a peak, reflecting a rapid spread of the disease. The blue curve with control measures in place increases more slowly and peaks at a lower level, suggesting that the interventions are operative in dropping the spread of the disease and the peak infectious population.

Together, these graphs highlight the influence of interventions in controlling the spread of an infectious disease, showing that appropriate measures can significantly reduce the number of exposed and infectious individuals in a population over time.

FIG. 12.

View large Download slide

(a) The outcome of control procedures on the quarantine and recovery of individuals during an infectious disease outbreak. (b) The outcome of control procedures on the quarantine and recovery of individuals during an infectious disease outbreak.

Figure 12(a) demonstrates the number of people in quarantine over time. The red curve indicates the situation without any control measures u₁ = 0, u₂ = 0, showing a slow increase in quarantined individuals. The blue curve represents the scenario with control measures u₁ ≠ 0, u₂ ≠ 0, which leads to a much more pronounced upsurge in the number of quarantined individuals. This implies that with the implementation of control measures, more people are being identified and quarantined to stop further spread of the disease.

Figure 12(b) displays the number of people who have recovered from the disease over time. Again, the red curve shows the recovery trajectory without control measures, which results in a slower and lower number of recoveries. The blue curve, with control measures in place, shows a more rapid and higher number of recoveries, indicating that the control measures are effective in both preventing the spread of the disease and helping more people recover.

These graphs suggest that the interventions, which could include quarantine, isolation, vaccination, or treatment strategies, not only reduce the spread of the infection but also contribute to a faster and more effective recovery of the population.

VII. CONCLUSION

Our study utilized a deterministic model for assessing the emergence and spread of COVID-19 cases, demonstrating its effectiveness. The positivity and boundedness of model (1) are determined. The reproduction number of model (1) is computed. Local and global stability of model (1) is also analyzed. By identifying the factors with the most significant impact on the basic reproduction number, such as the contact rate between susceptible and infectious individuals, our study highlights the crucial elements that need monitoring in pandemic response efforts. The sensitivity analysis conducted as part of our study is particularly valuable for understanding how different parameters interact and influence the spread of the virus. This can inform global public health policies, especially in regions with varying demographic and social characteristics. Our findings regarding the effectiveness of control measures such as social distancing, mask-wearing, and rapid quarantine are particularly relevant in the current global scenario. The use of control functions in our model to simulate these measures provides insights into their potential impact on dropping the spread of the virus. This aspect of our study can guide policymakers in implementing targeted interventions.

ACKNOWLEDGMENTS

Princess Nourah bint Abdulrahman University Researchers Supporting Project number (PNURSP2024R404), Princess Nourah bint Abdulrahman University, Riyadh, Saudi Arabia.

AUTHOR DECLARATIONS

Conflict of Interest

The authors have no conflicts to disclose.

Author Contributions

Kamil Shah: Conceptualization (equal); Data curation (equal); Investigation (equal); Methodology (equal); Supervision (equal); Validation (equal); Visualization (equal); Writing – original draft (equal); Writing – review & editing (equal). Jamal Shah: Conceptualization (equal); Data curation (equal); Investigation (equal); Methodology (equal); Supervision (equal); Validation (equal); Visualization (equal); Writing – original draft (equal); Writing – review & editing (equal). Ebenezer Bonyah: Conceptualization (equal); Data curation (equal); Investigation (equal); Methodology (equal); Supervision (equal); Validation (equal); Visualization (equal); Writing – original draft (equal); Writing – review & editing (equal). Tmader Alballa: Investigation (equal); Methodology (equal); Validation (equal); Visualization (equal); Writing – review & editing (equal). Hamiden Abd El-Wahed Khalifa: Data curation (equal); Validation (equal); Visualization (equal); Writing – review & editing (equal). Usman Khan: Methodology (equal); Writing – original draft (equal). Hameed Khan: Investigation (equal); Writing – original draft (equal).

DATA AVAILABILITY

The data that support the findings of this study are available from the corresponding author upon reasonable request.

REFERENCES

1

M.

Rojas

,

Y.

Rodríguez

,

D. M.

Monsalve

,

Y.

Acosta-Ampudia

,

B.

Camacho

,

J. E.

Gallo

,

A.

Rojas-Villarraga

et al, “

Convalescent plasma in Covid-19: Possible mechanisms of action

,”

Autoimmun. Rev.

19

(

7

),

102554

(

2020

).

https://doi.org/10.1016/j.autrev.2020.102554

Google Scholar

Crossref

PubMed

2

H. G.

Landau

, “

On some problems of random nets

,”

Bull. Math. Biophys.

14

,

203

–

212

(

1952

).

https://doi.org/10.1007/bf02477719

Google Scholar

Crossref

3

G.

Li

,

Y.

Fan

,

Y.

Lai

,

T.

Han

,

Z.

Li

,

P.

Zhou

,

P.

Pan

,

W.

Wang

,

D.

Hu

,

X.

Liu

,

Q.

Zhang

, and

J.

Wu

, “

Coronavirus infections and immune responses

,”

J. Med. Virol.

92

(

4

),

424

–

432

(

2020

).

https://doi.org/10.1002/jmv.25685

Google Scholar

Crossref

4

T. T.

Marinov

and

R. S.

Marinova

, “

Dynamics of COVID-19 using inverse problem for coefficient identification in SIR epidemic models

,”

Chaos, Solitons Fractals: X

5

,

100041

(

2020

).

https://doi.org/10.1016/j.csfx.2020.100041

Google Scholar

Crossref

5

L.

Matrajt

and

T.

Leung

, “

Evaluating the effectiveness of social distancing interventions to delay or flatten the epidemic curve of coronavirus disease

,”

Emerging Infect. Dis.

26

(

8

),

1740

–

1748

(

2020

).

https://doi.org/10.3201/eid2608.201093

Google Scholar

Crossref

6

C. E.

Mills

,

J. M.

Robins

, and

M.

Lipsitch

, “

Transmissibility of 1918 pandemic influenza

,”

Nature

432

(

7019

),

904

–

906

(

2004

).

https://doi.org/10.1038/nature03063

Google Scholar

Crossref

PubMed

7

R.

Jayatilaka

,

R.

Patel

,

M.

Brar

,

Y.

Tang

,

N. M.

Jisrawi

,

F.

Chishtie

,

J.

Drozd

, and

S. R.

Valluri

, “

A mathematical model of COVID-19 transmission

,”

Mater. Today: Proc.

54

,

101

–

112

(

2022

).

https://doi.org/10.1016/j.matpr.2021.11.480

Google Scholar

Crossref

PubMed

8

A. M.

Schiffer

, “

Controlling COVID-19

,”

Nat. Human Behav.

4

(

5

),

450

(

2020

).

https://doi.org/10.1038/s41562-020-0883-0

Google Scholar

Crossref

9

N. L.

Semendyaeva

,

M. V.

Orlov

,

T.

Rui

, and

Y.

Enping

, “

Analytical and numerical investigation of the SIR mathematical model

,”

Comput. Math. Model

33

(

3

),

284

–

299

(

2022

).

https://doi.org/10.1007/s10598-023-09572-7

Google Scholar

Crossref

10

A. D.

Algarni

,

A.

Ben Hamed

,

M.

Hamdi

,

H.

Elmannai

, and

S.

Meshoul

, “

Mathematical COVID-19 model with vaccination: A case study in Saudi Arabia

,”

PeerJ Comput. Sci.

8

,

e959

(

2022

).

https://doi.org/10.7717/peerj-cs.959

Google Scholar

Crossref

PubMed

11

K.

Shah

,

E. A.

Ismail

,

F. A.

Awwad

,

U.

Khan

,

J.

Shah

, and

E.

Bonyah

, “

Modeling the dynamics of co-infection between COVID-19 and tuberculosis with quarantine strategies: A mathematical approach

,”

AIP Adv.

14

(

7

),

075219

(

2024

).

https://doi.org/10.1063/5.0212904

Google Scholar

Crossref

12

S. A.

Carvalho

,

S. O.

da Silva

, and

I. D. C.

Charret

, “

Mathematical modeling of dengue epidemic: Control methods and vaccination strategies

,”

Theory Biosci.

138

(

2

),

223

–

239

(

2019

).

https://doi.org/10.1007/s12064-019-00273-7

Google Scholar

Crossref

PubMed

13

K.

Dietz

and

J. A. P.

Heesterbeek

, “

Daniel Bernoulli’s epidemiological model revisited

,”

Math. Biosci.

180

(

1–21

),

1

–

21

(

2002

).

https://doi.org/10.1016/s0025-5564(02)00122-0

Google Scholar

Crossref

PubMed

14

S.

Bhatter

,

S.

Bhatter

,

K.

Jangid

,

A.

Abidemi

,

K.

Owolabi

, and

S.

Purohit

, “

A new fractional mathematical model to study the impact of vaccination on COVID-19 outbreaks

,”

Decis. Anal. J.

6

,

100156

(

2023

).

https://doi.org/10.1016/j.dajour.2022.100156

Google Scholar

Crossref

15

S. S.

Negi

,

P. S.

Rana

,

N.

Sharma

, and

M. S.

Khatri

, “

A novel SEIAHR compartment model for accessing the impact of vaccination, intervention policies, and quarantine on the COVID-19 pandemic: A case study of most affected countries Brazil, India, Italy, and USA

,”

Comput. Appl. Math.

41

(

7

),

305

(

2022

).

https://doi.org/10.1007/s40314-022-01993-1

Google Scholar

Crossref

16

N.

Imai

,

A.

Cori

,

I.

Dorigatti

,

M.

Baguelin

,

C. A.

Donnelly

,

S.

Riley

, and

N. M.

Ferguson

, “

Report 3: Transmissibility of 2019-nCoV

,”

Imperial College London

625

,

1

–

6

(

2020

).

https://doi.org/10.25561/77148

Google Scholar

17

C. N.

Ngonghala

,

E.

Iboi

,

S.

Eikenberry

,

M.

Scotch

,

C. R.

MacIntyre

,

M. H.

Bonds

, and

A. B.

Gumel

, “

Mathematical assessment of the impact of non-pharmaceutical interventions on curtailing the 2019 novel Coronavirus

,”

Math. Biosci.

325

,

108364

(

2020

).

https://doi.org/10.1016/j.mbs.2020.108364

Google Scholar

Crossref

PubMed

18

F.

Ndaïrou

,

I.

Area

,

J. J.

Nieto

, and

D. F.

Torres

, “

Mathematical modeling of COVID-19 transmission dynamics with a case study of Wuhan

,”

Chaos, Solitons Fractals

135

,

109846

(

2020

).

https://doi.org/10.1016/j.chaos.2020.109846

Google Scholar

Crossref

PubMed

19

M.

Mandal

,

S.

Jana

,

S. K.

Nandi

,

A.

Khatua

,

S.

Adak

, and

T. K.

Kar

, “

A model based study on the dynamics of COVID-19: Prediction and control

,”

Chaos, Solitons Fractals

136

,

109889

(

2020

).

https://doi.org/10.1016/j.chaos.2020.109889

Google Scholar

Crossref

PubMed

20

D.

Prathumwan

,

K.

Trachoo

, and

I.

Chaiya

, “

Mathematical modeling for prediction dynamics of the coronavirus disease 2019 (COVID-19) pandemic, quarantine control measures

,”

Symmetry

12

(

9

),

1404

(

2020

).

https://doi.org/10.3390/sym12091404

Google Scholar

Crossref

21

D.

Okuonghae

and

A.

Omame

, “

Analysis of a mathematical model for COVID-19 population dynamics in Lagos, Nigeria

,”

Chaos, Solitons Fractals

139

,

110032

(

2020

).

https://doi.org/10.1016/j.chaos.2020.110032

Google Scholar

Crossref

PubMed

22

S.

Khajanchi

,

K.

Sarkar

, and

S.

Banerjee

, “

Modeling the dynamics of COVID-19 pandemic with implementation of intervention strategies

,”

Eur. Phys. J. Plus

137

(

1

),

129

(

2022

).

https://doi.org/10.1140/epjp/s13360-022-02347-w

Google Scholar

Crossref

PubMed

23

R. K.

Rai

,

S.

Khajanchi

,

P. K.

Tiwari

,

E.

Venturino

, and

A. K.

Misra

, “

Impact of social media advertisements on the transmission dynamics of COVID-19 pandemic in India

,”

J. Appl. Math. Comput.

68

(

1

),

19

–

44

(

2022

).

https://doi.org/10.1007/s12190-021-01507-y

Google Scholar

Crossref

24

M. A.

Acuña-Zegarra

,

S.

Díaz-Infante

,

D.

Baca-Carrasco

, and

D.

Olmos-Liceaga

, “

COVID-19 optimal vaccination policies: A modeling study on efficacy, natural and vaccine-induced immunity responses

,”

Math. Biosci.

337

,

108614

(

2021

).

https://doi.org/10.1016/j.mbs.2021.108614

Google Scholar

Crossref

PubMed

25

S.

Kurmi

and

U.

Chouhan

, “

A multicompartment mathematical model to study the dynamic behaviour of COVID-19 using vaccination as control parameter

,”

Nonlinear Dyn.

109

(

3

),

2185

–

2201

(

2022

).

https://doi.org/10.1007/s11071-022-07591-4

Google Scholar

Crossref

PubMed

26

A.

Venkatesh

and

M.

Ankamma Rao

, “

Mathematical model for COVID-19 pandemic with implementation of intervention strategies and cost-effectiveness analysis

,”

Results Control Optim.

14

,

100345

(

2024

).

https://doi.org/10.1016/j.rico.2023.100345

Google Scholar

Crossref

27

P.

van den Driessche

and

J.

Watmough

, “

Reproduction numbers and sub-threshold endemic equilibria for compartmental models of disease transmission

,”

Math. Biosci.

180

(

1–2

),

29

–

48

(

2002

).

https://doi.org/10.1016/s0025-5564(02)00108-6

Google Scholar

Crossref

PubMed

28

N.

Chitnis

,

J. M.

Hyman

, and

J. M.

Cushing

, “

Determining important parameters in the spread of malaria through the sensitivity analysis of a mathematical model

,”

Bull. Math. Biol.

70

,

1272

–

1296

(

2008

).

https://doi.org/10.1007/s11538-008-9299-0

Google Scholar

Crossref

PubMed

29

M. L.

Chambers

et al, “

The mathematical theory of optimal processes

,”

J. Oper. Res. Soc.

16

(

4

),

493

–

494

(

1965

).

https://doi.org/10.2307/3006724

Google Scholar

Crossref

30

L. S.

Pontryagin

,

Mathematical Theory of Optimal Processes

(

Springer

,

2023

).

Google Scholar

31

D. L.

Lukes

,

Differential Equations: Classical to Controlled

(

Springer

,

1982

), p.

337

.

Google Scholar

32

M. L.

Diagne

,

H.

Rwezaura

,

S. Y.

Tchoumi

, and

J. M.

Tchuenche

, “

A mathematical model of COVID-19 with vaccination and treatment

,”

Comput. Math. Methods Med.

2021

(

1

),

1250129

.

https://doi.org/10.1155/2021/1250129

Crossref

PubMed

33

L.

Wang

,

A. A.

Khan

,

S.

Ullah

,

N.

Haider

,

S. A.

AlQahtani

, and

A. B.

Saqib

, “

A rigorous theoretical and numerical analysis of a nonlinear reaction-diffusion epidemic model pertaining dynamics of COVID-19

,”

Sci. Rep.

14

(

1

),

7902

(

2024

).

https://doi.org/10.1038/s41598-024-56469-5

Google Scholar

Crossref

PubMed

34

T.

Hussain

,

M.

Ozair

,

F.

Ali

,

S. u.

Rehman

,

T. A.

Assiri

, and

E. E.

Mahmoud

, “

Sensitivity analysis and optimal control of COVID-19 dynamics based on SEIQR model

,”

Results Phys.

22

,

103956

(

2021

).

https://doi.org/10.1016/j.rinp.2021.103956

Google Scholar

Crossref

PubMed

Optimal control of COVID-19 through strategic mathematical modeling: Incorporating harmonic mean incident rate and vaccination

I. INTRODUCTION

II. MODEL FORMULATION

A. Positive characterization

1. Invariant region

B. Disease-free equilibria and reproduction number

III. LOCAL STABILITY OF DISEASE-FREE EQUILIBRIA

IV. SENSITIVITY ANALYSIS

V. NUMERICAL SIMULATION OF SVEIQR MODEL

VI. FORMULATION OF THE OPTIMAL CONTROL SVEIQR MODEL

A. An analysis of optimal control and its existence

B. Numerical simulation of optimal control

VII. CONCLUSION

ACKNOWLEDGMENTS

AUTHOR DECLARATIONS

Conflict of Interest

Author Contributions

DATA AVAILABILITY

REFERENCES

Citing articles via

Submit your article

Sign up for alerts

Resources

Explore

pubs.aip.org

Connect with AIP Publishing

Optimal control of COVID-19 through strategic mathematical modeling: Incorporating harmonic mean incident rate and vaccination

I. INTRODUCTION

II. MODEL FORMULATION

A. Positive characterization

1. Invariant region

B. Disease-free equilibria and reproduction number

III. LOCAL STABILITY OF DISEASE-FREE EQUILIBRIA

IV. SENSITIVITY ANALYSIS

V. NUMERICAL SIMULATION OF SVEIQR MODEL

VI. FORMULATION OF THE OPTIMAL CONTROL SVEIQR MODEL

A. An analysis of optimal control and its existence

B. Numerical simulation of optimal control

VII. CONCLUSION

ACKNOWLEDGMENTS

AUTHOR DECLARATIONS

Conflict of Interest

Author Contributions

DATA AVAILABILITY

REFERENCES

Citing articles via

Submit your article

Sign up for alerts

Related Content

Resources

Explore

pubs.aip.org

Connect with AIP Publishing

This Feature Is Available To Subscribers Only