🧬 biology

Optimal Control of SEIVZ Epidemic Model with Holling IV Incidence and Information Intervention

This paper proposes an optimal control strategy for an SEIVZ epidemic model incorporating Holling-IV incidence and information intervention, demonstrating that while the disease-free equilibrium is stable when $R_0 \le 1$ with a backward bifurcation at $R_0 = 1$ , the combined application of information and treatment controls effectively mitigates disease spread while minimizing associated costs.

Original authors: Lijuan Zhang, Fuchang Wang, Zhaolong Yuan, Xuqin Zhuang

Published 2026-07-06

📖 5 min read🧠 Deep dive

CC BY 4.0

Original authors: Lijuan Zhang, Fuchang Wang, Zhaolong Yuan, Xuqin Zhuang

Original paper licensed under CC BY 4.0 (https://creativecommons.org/licenses/by/4.0/). ⚕️ This is an AI-generated explanation of a preprint that has not been peer-reviewed. It is not medical advice. Do not make health decisions based on this content. Read full disclaimer

Imagine a bustling city where a new, tricky virus is spreading. This paper is like a sophisticated traffic control system designed to manage that virus, but with a twist: it accounts for how people's fear and knowledge change the way the virus moves, and it figures out the most cost-effective way to stop the outbreak.

Here is a breakdown of the paper's ideas using simple analogies:

1. The City Map: The SEIVZ Model

The authors built a digital map of the city, dividing the population into five groups, like different zones in a game:

S (Susceptible): People who haven't caught the virus yet. They are the "open gates."
E (Exposed): People who have caught the virus but are in "incubation." They have the virus but aren't showing symptoms yet, like a time bomb that hasn't gone off.
I (Infected): People who are sick and spreading the virus.
V (Vaccinated): People who have been vaccinated. They are like people wearing armor.
Z (Information): This is the unique part. It represents the "buzz" or "news" about the disease. It's not a person, but a measure of how much people know and worry about the virus.

2. The Traffic Rules: Non-Linear Spread

In old models, the virus spread like water flowing through a pipe: more sick people meant a straight line of more new infections.

The Paper's Twist: The authors say, "No, that's not how real life works." They use something called Holling Type IV incidence.
The Analogy: Imagine a crowded party. If a few people are coughing, everyone ignores it. But if half the room is coughing, people get scared, put on masks, and leave the party. The spread actually slows down because people are protecting themselves.
The model includes a "psychological brake." As the number of sick people ( $I$ ) gets high, the infection rate hits a ceiling because people start self-isolating.

3. The Two Levers: Information and Treatment

The paper asks: "How can the government stop this?" They propose pulling two specific levers:

Lever 1 ( $u_1$ ): Information Intervention. This is like a public service announcement campaign. The more you push this lever, the more people get the "Z" (Information) density up, making them more cautious.
Lever 2 ( $u_2$ ): Treatment. This is like sending in the medical teams to cure the sick.
The Catch: Both levers cost money. The goal isn't just to stop the virus; it's to stop it cheaply. The paper tries to find the "Goldilocks" strategy: enough information and treatment to win, but not so much that the budget explodes.

4. The Magic Formula: Optimal Control

To find the best strategy, the authors used a mathematical tool called Pontryagin's Maximum Principle.

The Analogy: Think of this as a GPS for the virus. The GPS calculates the perfect route to get from "Outbreak" to "Safe" while avoiding "Traffic Jams" (high costs).
The computer simulates thousands of scenarios to find the exact moment to turn the information dial up or the treatment dial up.
The Result: The simulations show that if you combine smart information campaigns with targeted treatment, you can flatten the curve of infections much faster than doing nothing or just doing one thing.

5. The Surprising Discovery: The "Backward Bifurcation"

This is the most technical but fascinating part of the paper.

The Rule of Thumb: Usually, if the "Reproduction Number" ( $R_0$ ) is less than 1, the disease dies out on its own. It's like a fire that runs out of fuel.
The Paper's Finding: Because of the complex way people react (the non-linear rules), the fire might not go out even if $R_0$ is less than 1.
The Analogy: Imagine a fire that is supposed to die out because there isn't enough wood. But, if the wind (information) changes direction just right, the fire can suddenly reignite and keep burning, even though the fuel supply looks low.
Why it matters: This means that just getting the numbers down to "safe" levels isn't always enough. You have to be careful, because the disease can hide and come back if the conditions (like how scared people are) shift.

6. The Simulation: The Test Drive

The authors ran a computer simulation (using a method called Runge-Kutta) to test their theory.

Scenario A (No Control): The virus spreads, peaks, and eventually fades, but it infects a lot of people.
Scenario B (With Control): When they applied the optimal mix of information and treatment, the peak of the virus was much lower, and the number of sick people dropped much faster.
The "Jitter": They found that the best control strategy isn't a smooth line. It's like a driver constantly tapping the brakes and gas pedal. The intensity of the information and treatment needs to change frequently based on how the virus is behaving at that exact moment.

Summary

This paper builds a smarter, more realistic model of how diseases spread by acknowledging that people react to fear. It proves that by carefully balancing public information and medical treatment, we can stop outbreaks more efficiently. However, it warns us that because human behavior is complex, a disease might not disappear even when the numbers look safe, requiring constant vigilance.

Technical Summary: Optimal Control of SEIVZ Epidemic Model with Holling IV Incidence and Information Intervention

Problem Statement
The paper addresses the limitations of existing infectious disease models, which often assume linear infection growth and fail to adequately incorporate the impact of information interventions and realistic incubation periods. Specifically, the authors note that the number of infected individuals does not increase linearly during transmission due to psychological factors (self-protection) and that medical resources in developing regions are often limited, leading to saturated treatment rates. Furthermore, while information interventions (e.g., public awareness campaigns) are known to influence disease dynamics, few studies integrate these non-linear incidence rates, incubation periods, and information-driven behavioral changes into a unified optimal control framework. The study aims to model these complexities to determine optimal strategies for minimizing infection spread and associated costs.

Methodology
The authors propose a deterministic SEIVZ (Susceptible-Exposed-Infected-Vaccinated-Information) compartmental model. The model incorporates the following features:

Non-linear Incidence: The infection rate follows a Holling Type IV functional response, $\frac{\beta SI}{1+bI^2}$ , where the term $\frac{1}{1+bI^2}$ represents the inhibition rate caused by susceptible individuals' self-protection behaviors when infection levels are high.
Saturated Treatment: Treatment is modeled using a saturated function, $\frac{cI}{1+\alpha I}$ , to reflect limited medical resources.
Information Dynamics: A variable $Z(t)$ represents information density, which grows based on infection levels and decays naturally. Information intervention influences the susceptible population's behavior.
Optimal Control Formulation: Two control variables are introduced: $u_1(t)$ representing the intensity of information intervention and $u_2(t)$ representing the intensity of treatment measures. The objective is to minimize an objective function $J(u_1, u_2)$ that balances the number of infected individuals against the costs of information transmission and treatment.
Analytical Tools: The study employs the next-generation matrix method to derive the basic reproduction number ( $R_0$ ). Stability analysis of disease-free and endemic equilibria is conducted. The existence of an optimal control solution is proven, and the optimal control strategy is characterized using the Hamiltonian function and Pontryagin's Maximum Principle.
Numerical Simulation: The system is solved numerically using the Runge-Kutta method to simulate various scenarios and validate theoretical findings.

Key Contributions

Model Development: The paper introduces a novel SEIVZ model that integrates Holling Type IV incidence, saturated treatment, latency periods, and a dynamic information intervention module.
Stability and Bifurcation Analysis: The authors calculate the basic reproduction number ( $R_0$ ) and prove the global asymptotic stability of the disease-free equilibrium when $R_0 \leq 1$ . They also identify conditions for a backward bifurcation at $R_0 = 1$ , indicating that disease persistence is possible even when $R_0 < 1$ under specific parameter conditions.
Optimal Control Strategy: The study derives the necessary conditions for optimal control using Pontryagin's Maximum Principle, providing explicit characterizations for the optimal information intervention and treatment intensities.
Numerical Verification: Simulations demonstrate that the proposed control strategies effectively reduce the number of infected individuals and that the inclusion of non-linear incidence leads to more complex, fluctuating control profiles compared to linear models.

Results

Equilibrium Stability: When $R_0 \leq 1$ , the disease-free equilibrium is both locally and globally asymptotically stable. When $R_0 > 1$ , the endemic equilibrium is globally asymptotically stable.
Backward Bifurcation: The analysis reveals a backward bifurcation at $R_0 = 1$ under specific conditions involving the interaction between information intervention rates, treatment saturation, and population parameters. This implies that reducing $R_0$ below 1 may not be sufficient to eradicate the disease if the infection rate remains above a certain threshold.
Control Effectiveness: Numerical simulations show that the combination of optimal control (information intervention and treatment) significantly reduces the peak number of infected individuals compared to scenarios with no control or no information intervention.
Control Variable Behavior: Under non-linear incidence, the optimal control variables ( $u_1, u_2$ ) exhibit frequent fluctuations and alternating patterns between 0 and 1, contrasting with the more continuous and stable controls observed in linear incidence models. This suggests that real-world strategies must be adjusted dynamically.
Impact of Information: Information intervention significantly increases the proportion of the vaccinated population and slows the decline of the susceptible population, though its direct impact on the number of infected patients is less pronounced than that of direct treatment in high-infection scenarios.

Significance and Claims
The paper claims that the proposed SEIVZ model offers a more realistic framework for understanding infectious disease dynamics by accounting for psychological self-protection, resource limitations, and information-driven behavioral changes. The authors assert that their findings highlight the potential of combining information intervention with optimal treatment strategies to mitigate disease spread.

The study modestly acknowledges several limitations:

The model simplifies real-world factors such as viral variants and heterogeneous population adherence.
Parameter values are based on existing literature rather than real-time data calibration.
The optimal control strategies are theoretical and do not account for practical implementation constraints like large-scale logistical costs.
The model does not currently incorporate age, occupation, or geographic heterogeneity, nor does it account for time delays in information dissemination.

The authors conclude that while the model provides useful insights into the dynamics of epidemics under information intervention, future work is needed to incorporate complex population structures, calibrate parameters with real-world data, and evaluate the practical feasibility of the proposed control strategies.