Abstract
The major objective of this paper is to examine a novel mean-reverting Ornstein–Uhlenbeck process-based stochastic SIRD model for transmission the epidemic disease that is a great crisis in numerous societies. For this purpose, the deterministic model is further converted into the stochastic form by allowing the infection rate satisfies the mean-reverting Ornstein–Uhlenbeck process to account the uncertainties involved in epidemic spread. At first using Lyapunov functions, the solution’s uniqueness and positivity will be demonstrated. Subsequently, the stochastic epidemic threshold \(\Re_{0}^{S}\) that controls the disease’s extinction and persistence in the mean is identified analytically. It has been established that when \(\Re_{0}^{S} < 1\) the disease will extinguish, whereas if \(\Re_{0}^{S} > 1\) the disease is persistent. At last, several numerical simulations are presented to demonstrate the findings of the hypothetical investigation results. These simulations served to vividly illustrate and validate the implications derived from the hypothetical analysis.
Similar content being viewed by others
Introduction
It would be fantastic for health authorities to be able to predict future outbreaks when an unknown disease first appears and starts to cause infections and fatalities. An important method for understanding how infectious diseases spread is the representation of infectious disease behaviors by dynamical systems1,2. It is presently thought that mathematical models have been crucial instruments in qualitative and quantitatively studying the transmission and management of infectious diseases3. Numerous academics used theoretical viewpoints and simulations performed numerically to analyze the transmission of various infectious diseases4,5,6,7.
The mathematical modeling of disease propagation is limited by various restrictions using the deterministic approach. Despite being easy to understand, they offer few details. As far as we know, environmental fluctuations influence the spread of infectious diseases and make it more complicated to foresee their behavior. In such cases, deterministic systems, while able to make very informative forecasts and previsions, are not appropriate enough8. So, there is a pressing need for a developed mathematical model that can take into account the randomness effect, especially in the context of a harmful infectious disease9. The uncertainty in a process’s progress is described by its stochastic model. Uncertainty is a result of randomness, which is a result of the evolution of the universe, and ignorance, which is a trait of humans.
In recent years, the stochastic perspective for modeling infectious diseases has received a lot of attention in research papers10,11,12,13,14. For example, Din and li presented a detailed analysis of a stochastic delayed model which governs the transmission mechanism of the Hepatitis B virus while considering the white noises and the effect of vaccinations15. A stochastic delayed VEIC epidemic model with a general incidence rate was considered in16. A stochastic hepatitis B model considering a time-delay in the transmission coefficient and immune response class was established by Din et al. in17. Nissar and Sabbar provided a new framework for modeling the dynamics of HIV/AIDS infection under antiretroviral therapy which aims to reduce a person's viral load to an undetectable level by tempered stable Lévy jumps9.
Several mathematicians have investigated further infectious epidemic models based on the SIR model, which is the simplest epidemic model and consists of three compartments18,19. The SIRD model20,21, SEIR model22,23,24, SEIS25, and MSEIR26 models are modifications of SIR that represent diverse epidemiological situations for diseases27,28,29. This paper proposes a new stochastic SIRD epidemic model that incorporates a mean-reverting Ornstein Uhlenbech process (MROU process). The stochastic version of this model incorporates randomness into the system, allowing for more realistic simulations of disease spread. Stochastic epidemic SIRD models have been used to investigate a variety of diseases including influenza, HIV/AIDS, Ebola, and COVID-19.
The Ornstein-Uhlenbech process (OU process) is the main structural element of the Barndroff-Nielsen-Shephard stochastic volatility model30,31. The most recent research has seen huge interest in modeling, which is based on MROU processes of the diffusion type. Larribi et al. investigated a new MROU-based stochastic SIRS epidemic model32. Wang et al. studied a SIS model with the OU process and earn the reproduction number of the model to determine the stochastic extinction and persistence3. A stochastic SIRC model based on the OU process was analyzed by Zhiming et al.33 Zhou and Shi proposed the stationary distribution and extinction of a stochastic SEIS epidemic model motivated by Black Karasinski process34.
Our new proposed model is designed to more accurately capture the complexities of infectious diseases by taking into account the changing behaviors of individuals over time. This study provides important insights into the dynamics of infectious diseases and highlights the need for more accurate modeling techniques to better predict and control the spread of outbreaks. It makes a significant contribution to the science of epidemiology and has significant implications for public health policy. The remaining sections of this article are structured as follows. The stochastic model’s problem formulation using the MROU process is presented in “Problem formulation”. This part demonstrates the existence and uniqueness of the system’s global positive solution. The necessary conditions for the disease’s extinction and persistence are created in “Disease extinction and persistence”. The mathematical simulation of the model utilizing simulated data is shown in “Numerical simulation”. “Conclusion” concludes with some discussion and conclusions.
Problem formulation
In this study, we introduce a MROU process-based stochastic SIRD epidemic model. The proposed model is designed to better capture the complex dynamics of infectious diseases. The SIRD epidemic model employed in this study is divided into four sections: susceptible people (S), infected people (I), recovered people (R), and deceased people (D). At time t, the associated variable for each compartment is denoted by the letters S(t), I(t), R(t), and D(t), respectively. The population size is calculated by adding these classes together.
The following ordinary differential equation system describes the traditional SIRD epidemic evolution20:
with [S(t0), I(t0), R(t0), D(t0)] as starting condition for the start time t0. The parameter \(\beta\) is the infection rate. The parameters \(\alpha ,\,\gamma\) are the recovered and death rate respectively. The infection rate of a disease (\(\beta\)), defined as the number of new infections for each unit of time, is well-known and significant in mathematical epidemiology. The assumption used by the majority of stochastic epidemic models is that the rate of infection in a random environment is a linear function of Gaussian white noise.
where the real-valued Brownian motion B(t) is defined on a probability space \((\Omega ,{\mathbb{F}},{\mathbb{F}}_{t} ,P)\) and σ is the standard deviation of the noise. The following perturbed model of (1) is suggested:
The other alternative model for \(\beta\) in a randomly varying environment, is the MROU process which has the following form:
where \(\theta ,\,\overline{\beta }\) and ξ are positive constants. The parameter \(\theta\) is the speed of revision. The parameter \(\overline{\beta }\) determine the long run mean level of the disease transmission rate \(\beta (t)\), and ξ is the intensity of the volatility. The results in references35 leads us to obtain the explicit solution as follows:
which \(\beta_{0} = \beta (0)\). It is clear that:
and
As is well known the term \(\xi \int\limits_{0}^{t} {\exp \left( { - \theta (t - s)} \right)dB(s) \sim N(0,\frac{{\xi^{2} }}{2\theta }} (1 - e^{ - 2\theta t} ))\). Then we have,
Equation (4) can be rewritten as follows:
where \(\sigma (t) = \frac{\xi }{{\sqrt {2\theta } }}\sqrt {1 - e^{ - 2\theta t} }\). Submitting (7) in to model (1) one can earn the following nonlinear SDE system:
Given the proper initial condition \(S(t_{0} ) = S_{0} > 0,\,I(t_{0} ) = I_{0} > 0,\,R(t_{0} ) = R_{0} \ge 0,\), and \(D(t_{0} ) = D_{0} \ge 0\).
To determine whether the solution is universal and beneficial, we must first investigate the dynamic behavior of an epidemic model. The following theorem will be utilized in this part to show that the model is correctly described and therefore biologically meaningful by ensuring that the solution stays in ∆.
Theorem 1
There exists a singular solution to system (8) for all t > 0 almost surely (a.s.) with any initial value.
Proof 1
Let us consider X (t) = (S(t), I(t), R(t), D(t)). The coefficients of the system (8) are locally Lipschitz continuous for all starting values, hence there exists a singular solution X(t) ∈ Γ on t ∈ [0, τe], which τe is the time of an explosion.
Now we prove that τe = ∞ almost surely (a.s.), and the solution is global. Let κ0 > 0 is large enough to allow each member of X(0) are all in interval \([\frac{1}{{\kappa_{0} }},\kappa_{0} ]\). Define,
For any κ ≥ κ0 and
Consider inf ϕ = ∞ where ϕ presents the empty set. Then τκ is increasing when κ → ∞. Set τ∞ = limκ→∞ τκ, then we derive that τ∞ ≤ τe a.s. It is clear that if τ∞ = ∞ a.s., it can be concluded that τe = ∞ a.s. then X (t) ∈ Γ for all t > 0. Let τ∞ ≠ ∞, then, there exists two constants \(\hat{\delta }\) > 0 and \(\tilde{\upsilon }\) ∈ (0, 1) such that P(τ∞ ≤ \(\hat{\delta }\)) ≥ \(\tilde{\upsilon }\). Then,
Consider the function \(\psi\): R+4 → R+, which is twice differentiable, with the definition below:
\(\psi\) is nonnegative function. Using the system (8) and Ito formula,
where
which is bounded and ℑ ∈ R+, then
and
The mathematical expectation is represented by E. Let Ωυ: = τe ≤ \(\hat{\delta }\) for υ ≥ υ1. From Eq. (9), we have \(P(\Omega_{\upsilon } ) \ge \tilde{\upsilon }\). Define
Then
Hence Eqs. (9) and (10) results:
which IΩτ is the indicator of set Ωτ. If υ → ∞ then ∞ > E(Γ0) + ℑ\(\hat{\delta }\) = ∞ which is contradiction. Then the hypothesis P(τ∞ ≤ \(\hat{\delta }\) ) > \(\tilde{\kappa }\) is incorrect and τ∞ = ∞ a.s.
Disease extinction and persistence
This section examines the factors that will determine whether the disease will disappear or continue to exist. According to the references32, the basic reproduction number of the relevant deterministic model of (8) is provided by:
This controls when an epidemic spreads or the disease just quietly disappears. Using this the threshold of stochastic model (8) is defined as follows36:
Theorem 2
Let X (t) = (S(t), I(t), R(t), D(t)) represent the solution to the system (8) with initial values X(0) ∈ ∆. If \(\Re_{{_{0} }}^{S} < 1\) or \(\frac{{\overline{\beta }^{2} \theta }}{\alpha + \gamma } < \xi^{2}\) then \(P(\mathop {\lim \sup }\limits_{t \to \infty } \frac{\log I(t)}{t} < 0) = 1\).
Specifically, the disease will almost surely become extinct exponentially.
Proof 2
We have using Ito formula,
where
and
We get the following results by integrating from 0 to t and dividing by t:
where \(\Phi (t) = \int\nolimits_{0}^{t} {\sigma (u)S(u)dB(u)}\).
Since f is increasing function on \((0,\frac{{2\overline{\beta }\theta }}{{\xi^{2} }})\), we earn:
If \(\Re_{0}^{S}\) < 1 then it implies that \(\mathop {\lim }\limits_{t \to \infty } \frac{{\int\limits_{0}^{t} {f(u)du} }}{t} = 0\).
However, we also have:
Which implies that \(\mathop {\lim }\limits_{t \to \infty } \frac{{\int\limits_{0}^{t} {g(S(u),u)du} }}{t} = 0\).
Finally, we have established that \(\Phi (t)\) is a local martingale and,
From references36 it can be established that \(\lim_{t \to \infty } \frac{\Phi (t)}{t} = 0\) then, \(\lim \sup_{t \to \infty } \frac{\log I(t)}{t} = 0\). Also, we have, \(f(S) = - \frac{{\xi^{2} }}{4\theta }(S - \frac{{2\overline{\beta }\theta }}{{\xi^{2} }})^{2} + \frac{{\overline{\beta }\theta }}{{\xi^{2} }} - (\alpha + \gamma ) \le \frac{{\overline{\beta }^{2} \theta }}{{\xi^{2} }} - (\alpha + \gamma )\).
Under condition of theorem, it is clear that,
From an epidemiological standpoint, it is more necessary to research the condition that contributes to the disease’s persistence in a community. Following this section, looks into the possibility that the disease will persist.
Theorem 3
For any initial value X(0) ∈ ∆ if \(\Re_{{_{0} }}^{S} > 1\), then \(\mathop {\lim \inf }\limits_{t \to \infty } I(t) = \infty\). It means that the disease will persist.
Proof 3
It is simple to demonstrate from Eq. (12) that:
From formula (11) we earn:
Integrating the two sides from 0 to t gives us,
Since \((\overline{\beta } + (\beta_{0} - \overline{\beta }))e^{ - \theta t} \le \overline{\beta } \vee \beta_{0}\), and from the first equation in system (8), the following inequality can be concluded:
\(dS \ge - (\overline{\beta } \vee \beta_{0} )NSdt - \sigma (t)S(t)I(t)dB(t)\).
Integrating between 0 and t, we earn:
Combining Eqs. (15) and (16) implies that,
where
For local martingales, the law of large numbers yields that \(\mathop {\lim }\limits_{t \to \infty } \frac{\Theta (t)}{t} = 0,\) a.s. then
Moreover \(f(N) = (\alpha + \gamma )(\Re_{0}^{S} - 1)\), then if \(\Re_{0}^{S} > 1\) we have:
Hence, \(\mathop {\lim }\limits_{t \to \infty } I(t) = \infty\).
Numerical simulation
A simulation study is carried out in this phase to evaluate the proposed model’s proficiency. An approximate diffusion system (8) can be used to demonstrate the analytical findings reported in the preceding sections. Here, we use the Euler approach for the simulation study. Three data sets corresponding to the model (8) with considering the different values of parameters are considered in Examples 1–3. In all examples the values of \(S(0) = 0.75,\,I(0) = 0.25,\,R(0) = 0,\,D(0) = 0,\,\beta_{0} = 0.15\), are chosen to be same.
Example 1
Generate the data sets from the model (8) with various parameter values as follows, \(\overline{\beta } = 0.22,\,\theta = 0.4,\,\xi = 0.25,\,\alpha = 0.1,\,\gamma = 0.1,\,\Re_{{^{0} }}^{S} = 0.904 < 1\) is obtained by straightforward calculation, and the condition of theorem 2 is tested. According to theorem 2 for the stochastic model, the disease will be removed from the population. Figure 1 shows the numerical simulation of this example and confirms the results of theorem 2. As we see, the disease will expire in about 50 days.
Example 2
Using the parameter values \(\overline{\beta } = 0.35,\,\theta = 1.3,\,\xi = 1.2,\,\alpha = 0.01,\,\gamma = 0.01\). It is easy to calculate that \(\Re_{{^{0} }}^{S} = 3.65 > 1\), then the condition of theorem 3 is verified. It means that the disease will persist. In Fig. 2 these results were be shown.
Example 3
This example considers the parameter values \(\overline{\beta } = 0.22,\,\theta = 0.13,\,\xi = 0.12,\,\alpha = 0.1,\,\gamma = 0.09\,\) that caused \(\Re_{{^{0} }}^{S} = 1\). Numerical simulations in Fig. 3 shows that the disease will be extinct in this critical situation. The scope of the authors’ next research activity is proof of this result.
Conclusion
This study proposes and evaluates a new stochastic SIRD epidemic model using an MROU process and a general non- linear incidence rate. In the suggested model, the existence and uniqueness of a positive global solution were verified. The stochastic system (8) has a threshold \(\Re_{0}^{S}\) that controls the extinction and persistence of epidemic diseases. In Theorem 2, we established that under a few additional conditions, when \(\Re_{0}^{S} < 1\) the disease disappears exponentially with probability one. Theorem 3 allowed us to demonstrate that the stochastic process I(t) is persistent in mean if \(\Re_{0}^{S} > 1\). The efficiency and correctness of the current work are demonstrated using numerical simulations with simulated data. The findings of Example 3 show that the disease will expire when the threshold \(\Re_{0}^{S} = 1\). Proof of this result is desirable for future research. The other strategy of this research is to investigate the extended Kalman filter for the stochastic SIRD model with the OU process in the future.
Data availability
The datasets generated and/or analyzed during the current study are available from the corresponding author upon reasonable request.
References
Babaei, A., Jafari, H., Banihashemi, S. & Ahmadi, M. Mathematical analysis of a stochastic model for spread of Coronavirus. Chaos Solitons Fract. 145 (2021).
Babaei, A., Ahmadi, M. & Jafari, H. A mathematical model to examine the effect of quarantine on the spread of coronavirus. Chaos Solitons Fract. 142 (2021).
Wanga, W., Cai, Y., Ding, Z. & Gu, Zh. A stochastic differential equation SIS epidemic model incorporating Ornstein Uh lenbeck process. Phys. A 509, 921–936 (2018).
Emvudu, Y., Bongor, D. & Koïna, R. Mathematical analysis of HIV/AIDS stochastic dynamic models. Appl. Math. Model. 40, 9131–9151 (2016).
Jajarmi, A., Ghanbari, B. & Baleanu, D. A new and efficient numerical method for the fractional modeling and optimal control of diabetes and tuberculosis co-existence. Chaos Interdiscip. J. Nonlinear Sci. (2019).
Khan, M. A., Hammouch, Z. & Baleanu, D. Modeling the dynamics of hepatitis E via the Caputo Fabrizio derivative. Math. Modell. Nat. Phen. 14 (2019).
Sajjadi, S., Baleanu, D., Jajarmi, A. & Mohammadi Pirouz, H. A new adaptive synchronization and hyper chaos control of a biological snap oscillator. Chaos Solitons Fract. 138 (2020).
Kiouach, D., Azami El-idrissi, S. E. & Sabbar, Y. An improvement of the extinction sufficient conditions for a higher-order stochastically disturbed AIDS/HIV model. Appl. Math. Comput. 447 (2023).
Nisar, K. S. & Sabbar, Y. Long-run analysis of a perturbed HIV/AIDS model with antiretroviral therapy and heavy-tailed increments performed by tempered stable Lévy jumps. Alex. Eng. J. 78, 498–516 (2023).
Ivorraa, B., Ferrández, M.R., Vela-Pérez a, M. & Ramos, A. M. Mathematical modeling of the spread of the coronavirus disease 2019 (COVID-19) taking into account the undetected infections: The case of China. Commun. Nonlinear Sci. Numer. Simul. (2020).
Tilahun, G. T., Demie, S. & Eyob, S. Stochastic model of measles transmission dynamics with double dose vaccination. Infect. Dis. Model. 1, 1 (2020).
Khoshnaw, S. H. A., Salih, R. H. & Sulaimany, S. Mathematical modeling for coronavirus disease in predicting future behaviors and sensitivity analysis. Math. Model. Nat. Phenom. 1, 1 (2020).
Sarkar, K., Khajanchi, S. & Nieto, J. J. Modeling and forecasting the COVID-19 pandemic in India. Chaos Solitons Fract 1, 1 (2020).
Maleki, M., Mahmoudi, M. R., Heydari, M. H. & Pho, K. H. Modeling and Forecasting the Spread and Death Rate of Coronavirus (COVID-19) in the World using Time Series Models. Chaos Solitons Fract. 1, 1 (2020).
Din, A. & Li, Y. Mathematical analysis of a new nonlinear stochastic hepatitis B epidemic model with vaccination effect and a case study. Eur. Phys. J. Plus 137, 1 (2022).
Din, A. The stochastic bifurcation analysis and stochastic delayed optimal control for epidemic model with general incidence function. Chaos 31, 1 (2021).
Din, A., Li, Y. & Yusuf, A. Delayed hepatitis B epidemic model with stochastic analysis. Chaos Solitons Fract. 1, 146 (2021).
Ariful, K. M., Kuga, K. & Tanimoto, J. Analysis of SIR epidemic model with information spreading of awareness. Chaos Solitons Fract. 1, 118–125 (2019).
Parsamanesh, M. & Farnoosh, R. On the global stability of the endemic state in an epidemic model with vaccination. Math. Sci. 1, 313–320 (2018).
Fanelli, D. & Piazza, F. Analysis and forecast of COVID-19 spreading in China, Italy and France. Chaos Solitons Fract. 1, 1 (2020).
Sebbagh, A. & Kechida, S. EKFSIRD model algorithm for predicting the coronavirus (COVID19) spreading dynamics (Rep, Sci, 2022). https://doi.org/10.1038/s41598-022-16496.
Jiao, J., Liu, Z. & Cai, Sh. Dynamics of an SEIR model with infectivity in incubation period and homestead-isolation on the susceptible. Appl. Math. Lett. 1, 1 (2020).
Aron, J. L. & Schwartz, I. B. Seasonality and period-doubling bifurcations in an epidemic model. J. Theoret. 1, 1 (1984).
Lu, G. & Lu, Z. Global asymptotic stability for the SEIRS models with varying total population size. Math. Bio sci. 1, 17–25 (2018).
Li, M. Y. & Muldowney, J. S. Global stability for the SEIR model in epidemiology. Math. Biosci. 125, 155164 (1995).
Asfour, H. A. & Ibrahim, M. On the differential fractional transformation method of MSEIR epidemic model. Int. J. Comput. Appl. 113, 1 (2015).
Dantas, E., Tosin, M. & Cunha, A. Jr. Calibration of a SEIRSEI epidemic model to describe the Zika virus outbreak in Brazil. Appl. Math. Comput. 338, 249–259 (2018).
Sun, G. Q., Xie, J. H., Huang, S. H., Jin, Z. & Li, M. T. Transmission dynamics of cholera: Mathematical modeling and control strategies. Commun. Nonlinear Sci. Numer. Simul. 45, 235–244 (2017).
Liu, X. & Stechlinski, P. Application of control strategies to a seasonal model of chikungunya disease. Appl. Math. Model. 39, 3194–3220 (2015).
Bandroff-Nielsen, O. E. & Shephard, N. Non-Gaussian OU based models and some of their uses in financial economics. J. R. Stat. Soc. Ser. B 63, 167–241 (2001).
Zhang, V. & Yuan, R. Pullback attractor for random chemostat model driven by colored noise. Appl. Math. Lett. 112, 1 (2021).
Laaribi, A., Boukanjime, B., Khalifi, M., Bouggar, D. & El Fatini, M. A generalized stochastic SIRS epidemic model incorporating mean-reverting OrnsteinUhlenbeck process. Phys. A J. 615, 1 (2023).
Zhiming, N., Daqing, J., Zhongwei, C. & Xiaojie, M. Analysis of Stochastic SIRC model with cross immunity based on Ornstein Uhlenbeck process. Qual. Theory Dyn. Syst. 22, 1 (2023).
Zhou, B. & Shi, N. Stationary distribution and extinction of a stochastic SEIS epidemic model motivated by Black Karasinski process. Appl. Math. Lett. 149, 1 (2024).
Nabati, P. & Hajrajabi, A. Three-factor mean reverting Ornstein-Uhlenbeck process with stochastic drift term innovations: Nonlinear autoregressive approach with dependent error. Filomat 36, 2345–2355 (2022).
Cai, S., Cai, Y. & Mao, X. A stochastic differential equation SIS epidemic model with two independent Brownian motions. J. Math. Anal. Appl. 474, 1536–1550 (2019).
Author information
Authors and Affiliations
Contributions
P.N., supervision, software, validation, writing—review and editing.
Corresponding author
Ethics declarations
Competing interests
The author declares no competing interests.
Additional information
Publisher's note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
Rights and permissions
Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article's Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article's Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creativecommons.org/licenses/by/4.0/.
About this article
Cite this article
Nabati, P. Introducing a novel mean-reverting Ornstein–Uhlenbeck process based stochastic epidemic model. Sci Rep 14, 1867 (2024). https://doi.org/10.1038/s41598-024-52335-6
Received:
Accepted:
Published:
DOI: https://doi.org/10.1038/s41598-024-52335-6
Comments
By submitting a comment you agree to abide by our Terms and Community Guidelines. If you find something abusive or that does not comply with our terms or guidelines please flag it as inappropriate.