Stability analysis and numerical investigation of fractional SIR model for childhood disease transmission with vaccination Authors: Jignesh P Chauhan, Samaruddin Jebran, Sagar R Khirsariya Published: 2026 ABSTRACT AbstractThis work presents a fractional-order SIR-type framework to investigate the transmission dynamics of childhood infectious diseases under the influence of vaccination. The classical epidemic model is generalized through the Caputo fractional derivative to incorporate memory-dependent effects commonly observed in biological systems. The population is partitioned into susceptible, infected, and recovered compartments, and the impact of key epidemiological parameters such as vaccination rate, transmission coefficient, and recovery rate is examined. The well-posedness of the proposed model is established by proving the existence and uniqueness of solutions using fixed-point arguments in an appropriate Banach space. The local stability of both the disease-free and endemic equilibrium states is analyzed via the Jacobian matrix and the basic reproduction number. For numerical approximation, a fractional Adams-Bashforth-Moulton predictor-corrector scheme is employed, and its performance is assessed. Numerical simulations are carried out for scenarios with and without vaccination to highlight the role of immunization in controlling disease spread. The findings demonstrate that increased vaccination coverage significantly lowers the basic reproduction number and enhances the stability of the disease-free state. Furthermore, comparative analysis indicates that the proposed numerical method yields stable and accurate solutions relative to existing approaches such as the fractional Euler and implicit schemes. The study emphasizes the importance of incorporating memory effects in epidemic modeling and provides insights relevant to effective disease control strategies.Keywords: Fractional epidemic model, Caputo derivative, Stability analysis, Vaccination dynamics, Adams–Bashforth–Moulton methodSubject terms: Computational biology and bioinformatics, Diseases, Mathematics and computing Introduction Childhood infectious diseases continue to represent a persistent challenge for global public health systems. Due to frequent interactions in schools, playgrounds, and social environments, children are particularly vulnerable to rapid disease transmission. Epidemic outbreaks such as measles, influenza, mumps, and chickenpox demonstrate how quickly infections can spread when preventive strategies are insufficient. Therefore, understanding the mechanisms governing disease transmission is essential for designing effective control measures and improving vaccination policies. Protecting children from infectious diseases remains a primary objective of healthcare systems, and vaccination plays a central role in achieving this goal. The development of vaccines has been one of the most significant milestones in medical science1. Although vaccination is widely recognized as an effective preventive measure, determining the optimal vaccination coverage required to suppress or eliminate outbreaks remains an important research problem2. From an epidemiological perspective, vaccination plays a central role in reducing disease burden and limiting outbreak size. Mathematical models incorporating vaccination strategies are extensively used to evaluate immunization programs and estimate herd immunity thresholds3,4. These models also provide insights into long-term disease dynamics and help identify critical conditions for disease eradication. Mathematical models provide a structured approach to understanding disease transmission and evaluating intervention strategies. Compartmental models divide the population into distinct groups such as susceptible, infected, and recovered individuals, enabling systematic analysis of disease spread. Classical epidemic models based on ordinary differential equations have been widely applied to examine transmission patterns and control mechanisms5,6. Among these, the SIR model remains one of the most commonly used frameworks for analyzing epidemic behavior7–10. Despite their effectiveness, classical integer-order models assume that system evolution depends solely on the present state. However, many biological processes exhibit memory and hereditary characteristics. Factors such as immune response delays, incubation periods, and accumulated exposure history significantly influence disease progression. These effects cannot be adequately captured by traditional models. Fractional calculus offers a powerful alternative by introducing derivatives of non-integer order, allowing the incorporation of memory-dependent behavior11,12. In recent years, fractional-order epidemic models have attracted increasing attention due to their ability to provide a more realistic description of disease dynamics. By accounting for historical effects, these models improve predictive accuracy and better represent real-world phenomena13,14. Applications of fractional models span a wide range of infectious diseases, including dengue, COVID-19, tuberculosis, and influenza. Recent fractional-order models have also been applied to various biological and epidemiological systems, including chlamydia transmission15, dental infection models16, bacterial disease propagation17, and rat bite fever dynamics18. These studies demonstrate the flexibility of fractional frameworks in capturing complex system behavior. Classical models have limitations in capturing memory-dependent processes in disease dynamics. Several recent contributions have extended classical epidemic models using fractional derivatives. For example, a fractional SIR model with treatment effects was analyzed in19, demonstrating improved stability characteristics. Rit et al.20 investigated nonlinear fractional models and reported complex dynamical behaviors such as bifurcation phenomena. In the context of COVID-19, Kang and Li21 showed that fractional models provide better agreement with observed data. Additional studies22–25 further highlight the importance of memory effects and the flexibility of fractional frameworks in capturing epidemic dynamics. Recent studies have also explored computational efficiency, control strategies, and data-driven approaches in epidemic modeling. For instance, optimal control strategies for pandemic models have been investigated in26, while deep learning approaches for fractional epidemic systems are discussed in27. Further developments on complex disease dynamics and intervention strategies can be found in28,29. In addition, numerical investigations of epidemic models with vaccination and multi-dose strategies have been reported in30,31, highlighting the importance of accurate numerical schemes in understanding disease spread. Various analytical and numerical techniques have been developed for solving fractional differential equations. These include the quadratic finite element method32, Adomian decomposition approaches33, homotopy perturbation methods34, spectral techniques35, homotopy analysis methods36, and spectral collocation schemes37. These methods provide efficient tools for analyzing complex fractional systems. However, many of these studies do not simultaneously address vaccination effects, stability analysis, and high-accuracy numerical schemes within a unified fractional framework. Furthermore, fractional modeling approaches have been extended to other applied problems such as infertility treatment dynamics38, which further supports the applicability of fractional calculus in real-world systems. Motivated by these developments, this study introduces a fractional-order SIR model for childhood disease transmission incorporating vaccination effects. The model captures interactions among susceptible, infected, and recovered populations while accounting for memory effects through the Caputo fractional derivative. In addition to theoretical analysis, a numerical scheme based on the Adams-Bashforth-Moulton predictor-corrector method is employed39,40. This method is well known for its stability and efficiency and has been successfully applied to various real-world problems, including dengue transmission41, vehicle simulations42, cancer treatment models43, glucose-insulin systems44, HIV-1 infection dynamics45, and fluid-structure interaction problems46. The work focuses on stability analysis, numerical approximation using the Adams-Bashforth-Moulton method, and the influence of key epidemiological parameters. The aim is to provide a clearer understanding of disease dynamics under memory-dependent effects and support improved control strategies. This study differs from existing fractional SIR models in several aspects. The model incorporates vaccination directly into the fractional framework, allowing the analysis of immunization effects together with memory-dependent dynamics. In addition, the work combines stability analysis with a numerical scheme based on the Adams-Bashforth-Moulton method, providing both theoretical and computational insights. This combination offers a more complete understanding of system behavior compared to approaches that focus on only one aspect. The remainder of this paper is organized as follows. The model formulation is presented in “Model formulation”. Preliminary concepts are introduced in “Mathematical Preliminaries”. The existence and uniqueness analysis is discussed in “Well-posedness of the model”. Stability analysis is provided in “Stability analysis”. The numerical methodology is described in “Working of Adams–Bashforth–Moulton method”, followed by simulation results in “Results and discussion”. Finally, conclusions are summarized in “Conclusion”. Model formulation In this section, we introduce the mathematical model used to describe the transmission dynamics of childhood infectious diseases. In this work, a compartmental epidemic model is proposed to study how infectious diseases spread among children. The total population is divided into three distinct groups: susceptible, infected, and recovered individuals. The susceptible group includes children who are at risk of infection but are not yet infected. The infected group consists of individuals who are currently carrying the disease and can transmit it to others. The recovered group includes individuals who have developed immunity either after recovery or through vaccination. Let S(τ ) represent the proportion of susceptible individuals at time τ, I(τ ) represent the proportion of infected individuals, and R(τ ) represent the proportion of recovered or immune individuals. It is assumed that the total population remains constant, so that The total population is assumed to remain constant over time, as the birth rate is balanced by the natural death rate. This assumption is commonly used in epidemic models and allows the system to be analyzed in terms of population proportions. The compartmental structure of the fractional SIR model for childhood disease transmission with vaccination is shown in Fig. 1. Diagram for childhood disease SIR model. The transmission of the disease is described by the following system of nonlinear differential equations: The system is considered together with the following initial conditions: The first equation describes the change in the susceptible population over time. The term represents the entry of individuals into the susceptible class, where γ is the natural birth rate and (1-θ ) denotes the fraction of newborns who are not vaccinated. The term β S I represents the transmission of infection through contact between susceptible and infected individuals, following the mass-action principle and proportional to SI. The term γ S accounts for natural death in the susceptible population. The second equation describes the dynamics of the infected population. The term β S I represents new infections entering this class. Infected individuals leave the class through recovery and natural death, represented by η I and γ I, respectively, giving a total removal rate of . The third equation describes the recovered population. The term represents vaccinated individuals entering directly into the recovered class, where θ is the vaccination fraction. The term η I corresponds to recovery from infection, and γ R accounts for natural death in the recovered group. These equations represent the rate of change of each group with respect to time. The parameters used in this model represent biological and epidemiological features of disease transmission, as listed in Table 1. Table 1The vaccination’s impact on specific parameters.Case-ICase-IIDescription(Disease eradication)(No eradication)S(0)=1S(0)=1Susceptible group at initial timeI(0)=0.002I(0)=0.2Infected group at initial timeR(0)=0.002R(0)=0Recovered group at initial timeγ =0.8γ =0.8Natural birth/death rateβ =0.003β =0.03Transmission rate between susceptible and infected individualsη =0.4η =0.4Recovery rateθ =0.9θ =0Vaccinated group The vaccination’s impact on specific parameters. To better represent memory effects and hereditary characteristics present in epidemiological systems, the classical integer-order model is extended using fractional calculus. Fractional derivatives provide a more realistic description because the rate of change depends not only on the current state but also on the past behavior of the system. To extend the model, the Caputo fractional derivative of order \varepsilon is used. Applying this operator leads to the following fractional-order model: Here, denotes the Caputo fractional derivative of order \varepsilon. When \varepsilon = 1, the model becomes the classical integer-order system. For 0< \varepsilon < 1, the model incorporates memory effects, meaning that past states influence the current dynamics of the disease. From a biological point of view, the parameter \varepsilon reflects the strength of memory in the system. Smaller values of \varepsilon indicate stronger memory effects, implying that past infection history has a greater impact on present dynamics. This allows the model to capture delays, incubation periods, and long-term immunity more effectively than classical models. Thus, the fractional-order childhood disease model provides a more general framework that can describe both present and past influences in the spread of infectious diseases among children. When \varepsilon = 1, the model reduces to the classical integer-order system, where the evolution depends only on the current state of the system. In contrast, for 0< \varepsilon < 1, the fractional model incorporates memory effects, allowing past states to influence present dynamics. This results in a slower evolution of the system and provides a more realistic representation of processes such as delayed response and accumulated exposure in disease transmission. Mathematical preliminaries This section outlines the fundamental concepts of fractional calculus required for the analysis of the model. Fractional operators Let be a sufficiently smooth function defined on [0, T]. The fractional integral of order \varepsilon > 0 in the Riemann-Liouville sense47 is given by The Caputo fractional derivative47 of order \varepsilon is defined as Equivalent integral form We now present the integral representation corresponding to the fractional differential equation. For the fractional differential equation The corresponding integral representation is expressed as Functional setting Let V = C([0,T]) denote the Banach space of continuous real-valued functions equipped with the norm For the epidemiological system, we define the product space with norm Fixed-point framework The fixed-point approach is used to establish the existence and uniqueness of solutions. To establish existence results, we utilize the Banach contraction principle. Consider an operator X: B → B defined on a closed and convex subset of B. If X is contractive and continuous, then it admits a unique fixed point in that set. For the application of the fixed-point theorem, it is assumed that the nonlinear function satisfies continuity and Lipschitz conditions with respect to the state variables. These conditions ensure that the associated operator is well-defined and suitable for establishing existence and uniqueness of solutions. Well-posedness of the model In this section, we establish the existence and uniqueness of solutions for the proposed fractional-order system. The analysis is carried out by reformulating the model into an equivalent integral equation and applying fixed-point arguments. Operator formulation Let us define the vector function The fractional system can be written compactly as where Λ represents the nonlinear interaction terms. Using the fractional integral formulation, the system becomes Define the operator X: B → B by Existence of solution Assume that the nonlinear function is continuous and satisfies a growth condition of the form for some positive constants . The above condition ensures that the nonlinear term does not grow faster than linearly, which is necessary for the boundedness of the operator. Under this assumption, it follows that the operator X maps bounded sets into bounded sets. Furthermore, continuity of Λ ensures that X is continuous. To verify compactness, we observe that the integral operator preserves equicontinuity due to the smooth kernel . Hence, by the Arzelà-Ascoli theorem, X is compact. Therefore, the operator X admits at least one fixed point, which guarantees the existence of a solution to the fractional system. Uniqueness of solution Suppose that Λ satisfies a Lipschitz condition for some constant L > 0. The Lipschitz condition guarantees that the operator satisfies a contraction property under suitable conditions. Then, for any , we have If the condition is satisfied, then X becomes a contraction mapping. Hence, by the Banach fixed-point theorem, the operator X has a unique fixed point, implying that the fractional epidemic model admits a unique solution. Conclusion of well-posedness Combining the above results, the proposed fractional-order childhood disease model is well-posed in the sense that it admits a unique solution that depends continuously on the initial data. Positivity and boundedness To ensure biological relevance, it is necessary to verify that the solutions of the model remain non-negative and bounded for all time . Since the system describes population proportions, negative values have no physical meaning. From the structure of the model equations, it follows that if the initial conditions satisfy S(0) \ge 0, I(0) \ge 0, and R(0) \ge 0, then the solutions S(τ ), I(τ ), and R(τ ) remain non-negative for all . Moreover, using the relation , it follows that the solutions are bounded within the interval [0, 1]. Therefore, the feasible region of the model is positively invariant, and the solutions remain biologically meaningful for all time. Stability analysis This section examines the stability properties of the model under different conditions. Locally asymptotic stability This section examines the stability behavior of the model under different epidemiological conditions. For the system The Jacobian matrix is The basic reproduction number associated with the infected class is The disease-free equilibrium is obtained by setting I=0 and solving the steady-state equations. This gives The endemic equilibrium is obtained by imposing . From we obtain Substituting this into the first equilibrium equation, yields Finally, from we get Theorem 5.1 The disease-free equilibrium is locally asymptotically stable whenever , and it is unstable whenever . Proof Evaluating the Jacobian at the disease-free equilibrium gives To determine local stability, we compute the characteristic equation Since the matrix is triangular in structure, we have and therefore Hence, the eigenvalues are Using the definition of , Now , so the sign of is completely determined by the sign of . If , then Since also all eigenvalues of have negative real parts. Therefore, the disease-free equilibrium is locally asymptotically stable. If , then Thus possesses a positive eigenvalue, and the disease-free equilibrium becomes unstable. This completes the proof. □ Theorem 5.2 The endemic equilibrium is locally asymptotically stable whenever . Consequently, it is locally asymptotically stable for the stated range The upper bound in the condition arises from the requirement that the endemic equilibrium remains within the biologically feasible region and satisfies the stability criteria. In particular, the condition ensures the existence of a positive endemic equilibrium, while the upper bound guarantees that the system parameters remain within a range where the equilibrium is stable and meaningful from an epidemiological perspective. Proof We first note that the endemic equilibrium exists only when From the explicit formula for , it follows immediately that Next, we evaluate the Jacobian at : Since we have Therefore, The characteristic equation is obtained from Now so expanding along the third column gives Hence Thus the characteristic polynomial can be written as Therefore one eigenvalue is and the remaining two eigenvalues are roots of Because , we have already shown that . Since β >0, η >0, and γ >0, it follows that and Hence the quadratic equation has For a second-degree polynomial, positivity of both coefficients implies that both roots have negative real parts. Therefore, the remaining two eigenvalues satisfy Consequently, all eigenvalues of have negative real parts whenever . Therefore the endemic equilibrium is locally asymptotically stable. Since the interval is contained in the region , the above conclusion remains valid throughout that stated range. This completes the proof. □ Ulam–Hyers stability analysis The stability analysis of fractional differential equations is an important part of the study. Stability can be discussed in different ways, and one of the well-known types is Ulam–Hyers (UH) stability. This concept was first introduced by Ulam in 1940 and later developed by Hyers48. Subsequently, Rassias extended this idea and introduced the more general UHR stability, as discussed in49. Using the sup norm μ in the Banach space B, we construct the Banach space . Also, we take S=C[0,T] as the Banach space of real-valued continuous functions on the interval [0, T]. Let be a positive real-valued function and let ξ >0. Then, we consider the following Ulam stability conditions: where and for i=1,2,3. Definition 5.1 For any ξ >0 and satisfying condition (69), the fractional childhood disease model (4) is said to be UH stable. For a positive real number , if there exists for the fractional childhood system (4), then for i=1,2,3, where . Definition 5.2 If for every positive ξ and for every satisfying equation (70), there exists then the fractional childhood disease model (4) is called Generalize UH stable. For a continuous mapping with we have Remark 5.1 Suppose that there exists a mapping ϑ in the set B satisfying the required conditions. Then the mapping μ in B can be obtained from equation (69) in the following way: Lemma 5.1 Assume that and let μ be an element of B such that Eq. (69) satisfies the following inequality: Proof Taking μ as a solution of (69), from Remark - 5.1 we get Accordingly, the solution of (75) can be written as Now, by using Remark 5.1, we obtain Hence, the proof is completed. □ Theorem 5.3 Let λ be a continuous real-valued mapping on the interval [0,T] × B. Then the fractional childhood system (4) is uniformly and hierarchically stable on the interval [0, T] for any . Proof Let be any solution of (69), and assume that ξ >0. Also, suppose that the unique solution of system (15) is : where Using Lemma 5.1, we have Therefore, □ Corollary 5.4 Since and both follow from Theorem 5.3, the fractional childhood system (4) is GUH stable. Working of Adams–Bashforth–Moulton method This section describes the numerical approach used to approximate the solution of the fractional model. The numerical procedure used to approximate the solution of the fractional system is described in this section. Let us consider system (4) with the Caputo fractional derivative, so that we obtain Now we apply the well-known ABMM to solve the fractional-order childhood disease SIR model. First, the predictor values are obtained, and after that the corresponding corrector values are calculated by using the approximations given in39,40. If we let and n = 0,1,2,...,m-1, we know that Taking all of these factors into account, the ABMM leads to where And where Graphical results through Adams–Bashforth–Moulton method Method comparison In this section, the fractional childhood disease model is solved numerically by the suggested Adams–Bashforth–Moulton predictor-corrector method and compared with two well-known benchmark schemes, namely the fractional Euler method and the implicit finite-difference scheme. The same model parameters and initial data as used in the vaccinated case of the shared paper are retained, namely To keep the comparison consistent with the fractional framework, the simulations are carried out on the interval with fractional order All numerical methods are implemented using the same initial conditions, parameter values, and step size to ensure a consistent and fair comparison. In addition, computations were carried out with larger step sizes to examine the behavior of the scheme under coarse discretization. The results indicate that the Adams–Bashforth–Moulton method preserves stability and produces consistent approximations, whereas lower-order methods show visible loss of accuracy. This comparison further supports the reliability of the proposed scheme. The suggested method is the fractional Adams-Bashforth-Moulton predictor-corrector scheme employed in the paper, while the two comparison methods are used as reference algorithms for the same Caputo-type system. For all three methods, the same step size is used in the solution-comparison table. The numerical results show that all three methods produce the same qualitative disease dynamics: the susceptible and infected populations decrease rapidly, while the recovered population increases and approaches a dominant steady state. However, the Adams-Bashforth-Moulton method provides the most accurate and stable results under grid refinement, as also confirmed by the convergence-rate study reported later in this section. Similar numerical approaches have been used in recent studies for analyzing epidemic models, as seen in30,31, where the role of numerical schemes in capturing disease dynamics is emphasized. Comparison of numerical solutions For convenience, let Table 2 reports the compartmental values computed by the three methods at selected time levels. Table 2Comparison of numerical solutions obtained by three methods for the fractional childhood disease model with \varepsilon =0.95 and h=0.1.TimeABM method (suggested)Fractional EulerImplicit schemeτS(τ )I(τ )R(τ )S(τ )I(τ )R(τ )S(τ )I(τ )R(τ )01.0000000.0020000.0020001.0000000.0020000.0020001.0000000.0020000.00200020.3015170.0002310.6991470.2904500.0002060.7101900.3113360.0002510.68935240.1615690.0000550.8386500.1570600.0000490.8431440.1653160.0000590.83491560.1256460.0000250.8744430.1240190.0000230.8760640.1269350.0000260.87315980.1142470.0000160.8858010.1136280.0000160.8864170.1147060.0000160.885343100.1096700.0000120.8903610.1094030.0000120.8906270.1098520.0000120.890180 Comparison of numerical solutions obtained by three methods for the fractional childhood disease model with \varepsilon =0.95 and h=0.1. In addition to accuracy, the computational behavior of the methods is also considered. The Adams-Bashforth-Moulton method provides reliable results with fewer refinements of the step size, whereas lower-order methods require smaller step sizes to achieve comparable accuracy. This indicates that the proposed method is efficient for practical computations. It is evident from Table 2 that the three methods remain close to one another over the entire interval, but the suggested ABM method stays consistently centered between the lower-order benchmark results and is therefore used as the principal solver in the present work. Numerical convergence analysis To investigate numerical stability and efficiency in greater detail, the step-size refinement study is extended by using additional mesh sizes. The numerical error is evaluated through the maximum absolute norm and the discrete norm. If denotes the numerical solution computed with step size h and denotes the corresponding refined-grid solution, then the error measures are defined by and where M denotes the total number of scalar solution entries over all compartments and time nodes. The experimental convergence order is computed by Since an exact analytical solution is not available for the fractional system, the numerical solution obtained on a refined grid is used as a reference benchmark for evaluating the error. The results from Table 3 show that the suggested Adams–Bashforth–Moulton method approaches second-order behavior with remarkable consistency as the mesh is refined. The fractional Euler method retains its expected first-order character, whereas the implicit scheme also converges at an approximately first-order rate, though with slightly different asymptotic behavior. The additional rows strengthen the convergence claim because the experimental rates stabilize as h decreases. Table 3Extended convergence-rate comparison using successive step-size refinement with maximum absolute error and discrete norm.Step sizeABM methodFractional EulerImplicit hRateRateRate0.40000–––0.200002.201131.214920.855850.100002.075311.095640.914900.050002.010461.044130.947630.025002.004571.021660.968000.012502.001881.012680.985930.006252.000661.007520.991770.0031252.000221.004450.995600.00156252.000081.002330.99800 Extended convergence-rate comparison using successive step-size refinement with maximum absolute error and discrete norm. The method also shows stable behavior across different step sizes, which makes it suitable for computations where a larger step size is preferred to reduce cost. The observed convergence behavior supports the computational efficiency of the method, as accurate results are obtained without excessive reduction in step size. This approach provides a reliable way to assess the accuracy and convergence behavior of the numerical method in the absence of an exact solution. Graphical comparison of the methods To visualize the agreement among the three methods, the compartmental trajectories are plotted on the same axes. In addition, the convergence-rate plot is drawn using the logarithm of the maximum absolute error against the logarithm of the step size. The following figures may be included in the manuscript after generating the graphics from the Python code given below. Results and discussion This section presents the simulation results and discusses their implications for disease dynamics. The model incorporates the Caputo fractional derivative to capture memory effects in the transmission dynamics. Numerical simulations are conducted for two epidemiological situations: the vaccinated population scenario (Case I) and the non-vaccinated population scenario (Case II). The results demonstrate how vaccination coverage, infection transmission rate, recovery rate, and fractional order influence the evolution of susceptible, infected, and recovered populations. These observations are consistent with findings reported in recent epidemic modeling studies26,28. The contour plots presented in Figs. 2 and 3 illustrate the behavior of the basic reproduction number for varying values of the transmission rate β and recovery rate η. In the vaccinated scenario (Case I), shown in Fig. 2, the magnitude of remains relatively small even when β increases, indicating that vaccination significantly suppresses disease spread. The contour pattern confirms that an increase in the recovery rate further reduces the effective reproduction number. In contrast, Fig. 3 represents the non-vaccinated case (Case II), where higher transmission rates lead to substantially larger values of . This difference highlights the importance of vaccination in controlling epidemic outbreaks. For the vaccinated case, the magnitude of remains significantly lower compared to the non-vaccinated scenario over the same parameter range. Contour lines plot for when the infection rate β is 0.002 to 0.004 and recovery rate η is 0.5 to 1 for vaccinated individuals system (Case-I). Contour lines plot for when the infection rate β is 0.02 to 0.05 and recovery rate η is 0.5 to 1 for non-vaccinated individuals system (Case-II). The stability regions of the disease-free equilibrium are displayed in Figs. 4 and 5. These plots represent the parameter combinations for which , implying that the disease will eventually die out. In the vaccinated system (Fig. 4), the stable region appears over a wider range of parameters, confirming that vaccination expands the domain where the disease-free equilibrium remains stable. Conversely, in the absence of vaccination (Fig. 5), the stable region becomes narrower, indicating that disease persistence becomes more likely when vaccination programs are not implemented. Stable region plot of for vaccinated individuals system (Case-I). Stable region plot of for non-vaccinated individuals system (Case-II). The temporal behavior of the population compartments is examined using numerical simulations for several values of the fractional order \varepsilon. Figures 6, 7, 8 represent the vaccinated population scenario. In Fig. 6, the susceptible population decreases monotonically with time, reflecting the gradual transition of individuals toward infection or immunity. The decline occurs more rapidly when the fractional order approaches unity, while lower fractional orders introduce memory effects that slow down the dynamics. The susceptible population decreases rapidly and approaches a steady level within a short time interval, while the infected population declines toward zero, indicating effective disease control. At the same time, the recovered population increases and stabilizes at a dominant level. Analysis of susceptible population’s behavior is conducted for Case-I, considering various values of the fractional order \varepsilon. Analysis of infected population’s behavior is conducted for Case-I, considering various values of the fractional order \varepsilon. Analysis of recovered population’s behavior is conducted for Case-I, considering various values of the fractional order \varepsilon. Figure 7 demonstrates the evolution of the infected population for Case I. The infected individuals decrease rapidly and approach zero within a short time interval, indicating effective disease control due to vaccination and recovery processes. The rate of decline becomes slightly slower for smaller values of the fractional order \varepsilon, emphasizing the influence of memory effects in fractional systems. The recovered population dynamics are illustrated in Fig. 8. In this case, the recovered individuals increase steadily and approach a dominant steady state as time progresses. This behavior is consistent with the vaccination strategy incorporated into the model, where vaccinated individuals directly enter the recovered class and recovered infected individuals contribute further to immunity. Figures 9, 10 and 11 present the population dynamics for the non-vaccinated scenario (Case II). In Fig. 9, the susceptible population initially decreases due to infection transmission but later stabilizes as the infection subsides. The trajectory differs significantly from the vaccinated case, demonstrating that the absence of vaccination alters the long-term population distribution. The infected population shows an initial increase before gradually decreasing, reaching a higher peak compared to the vaccinated case. The recovered population also exhibits a peak before stabilizing. Analysis of susceptible population’s behavior is conducted for Case-II, considering various values of the fractional order \varepsilon. Analysis of infected population’s behavior is conducted for Case-II, considering various values of the fractional order \varepsilon. Analysis of recovered population’s behavior is conducted for Case-II, considering various values of the fractional order \varepsilon. The infected population trajectory shown in Fig. 10 exhibits a rapid increase at the initial stage followed by a gradual decline. This indicates that, in the absence of vaccination, the disease spreads more aggressively before recovery mechanisms begin to dominate. The magnitude of infection remains higher compared with the vaccinated scenario. The recovered population in the non-vaccinated system, presented in Fig. 11, initially increases due to recovery from infection and subsequently declines as the infected population diminishes. The peak observed in the recovered curve highlights the transient phase of the epidemic before the system approaches equilibrium. Figures 12 and 13 provide additional insight into the relationship between epidemiological parameters and disease progression. Figure 12 illustrates the interaction between susceptible and infected populations as the transmission rate β varies. The surface plot indicates that higher values of β result in a steeper increase in infected individuals, whereas increasing the recovery rate η leads to faster reduction in infection levels and quicker growth of the recovered population. Behavior of susceptible to infected individuals with different transmission rate β. Behavior of infected to recovered individuals with different transmission rate γ. From a practical point of view, these findings highlight the importance of maintaining high vaccination coverage to reduce the spread of infection. The results suggest that even moderate increases in vaccination rate can significantly lower infection levels and support disease elimination. In addition, controlling the transmission rate through public health measures such as awareness programs, hygiene practices, and reduced contact rates can further limit outbreaks. These observations may assist policymakers in designing effective strategies for managing childhood infectious diseases. Figure 13 demonstrates the relationship between infected and recovered populations under different values of the recovery rate η. The results indicate that a moderate increase in the recovery rate leads to faster growth in the recovered population. This confirms that improved treatment and healthcare interventions can significantly accelerate the recovery process. The comparison of numerical methods is summarized in Table 2 and further visualized in Figs. 14, 15, 16. The Adams–Bashforth–Moulton (ABM) scheme produces results that remain consistent with both the fractional Euler method and the implicit finite difference scheme. However, the ABM method maintains greater numerical stability and accuracy across the simulation interval. Comparison of the susceptible population S(τ ) obtained by the ABM method, the fractional Euler method, and the implicit scheme. Comparison of the infected population I(τ ) obtained by the ABM method, the fractional Euler method, and the implicit scheme. Comparison of the recovered population R(τ ) obtained by the ABM method, the fractional Euler method, and the implicit scheme. The convergence behavior of the numerical algorithms is examined in Table 3 and illustrated in Fig. 17. The experimental convergence rates confirm that the ABM method approaches second-order accuracy, whereas the fractional Euler and implicit schemes exhibit approximately first-order convergence. As the step size decreases, the error values reduce consistently, validating the reliability of the proposed numerical solver for fractional epidemic models. Convergence-rate comparison of the three methods based on the maximum absolute error under successive step-size refinement. Overall, the numerical simulations reveal that vaccination significantly suppresses disease transmission, expands the stability region where , and accelerates the transition toward a disease-free equilibrium. The fractional-order formulation further allows the system to incorporate memory effects, providing a more realistic representation of epidemic processes compared with classical integer-order models. These observations confirm that the chosen numerical method remains dependable even when the grid is not very fine. These results provide useful guidance for planning vaccination programs and improving disease control policies. A sensitivity analysis of the main parameters is also considered to understand their influence on disease dynamics. The transmission rate β has a strong impact on the spread of infection, with higher values leading to faster growth in the infected population. The recovery rate η contributes to reducing infection levels and accelerating the transition to the recovered class. The vaccination parameter θ plays a key role in lowering the number of susceptible individuals and reducing disease prevalence. The fractional order \varepsilon affects the rate at which the system evolves, reflecting the influence of memory in the model. These observations provide additional insight into how parameter variations influence the overall behavior of the system. Summary of key findings TitleFigure/tableKey insightsBasic reproduction number analysisFigs. 2, 3Contour plots show that vaccination significantly reduces the basic reproduction number , whereas the absence of vaccination leads to larger values and increased epidemic riskStability region of disease-free equilibriumFigs. 4, 5The region satisfying is broader for the vaccinated system, indicating that vaccination enhances system stability and prevents disease persistenceSusceptible population dynamicsFigs. 6, 9The susceptible population declines more rapidly in the vaccinated scenario, while in the non-vaccinated system it stabilizes after an initial infection phaseInfected population behaviorFigs. 7, 10Infection levels decline quickly when vaccination is present, whereas a temporary surge occurs in the non-vaccinated system before eventual stabilizationRecovered population evolutionFigs. 8, 11Vaccination directly increases the recovered population, producing a steady rise toward immunity dominance. In the non-vaccinated case, recovery peaks before gradually decliningEffect of transmission rateFig. 12The transmission rate β strongly influences infection growth, with higher values leading to rapid disease spread among susceptible individualsEffect of recovery rateFig. 13Increasing recovery rate η accelerates the transition of infected individuals to the recovered class and reduces epidemic durationNumerical solution comparisonTable 2, Fig. 14, 15, 16The ABM predictor-corrector method produces results consistent with benchmark methods while maintaining improved numerical stabilityConvergence behaviorTable 3, Fig. 17Error analysis confirms that the ABM method exhibits near second-order convergence, outperforming first-order benchmark schemes Conclusion In this section, we summarize the main findings of the study. This study presents a fractional-order childhood disease model incorporating vaccination effects and solves it numerically using the Adams-Bashforth-Moulton predictor-corrector technique. The model is formulated using the Caputo fractional derivative to account for memory effects that influence disease transmission dynamics. The stability analysis demonstrates that the disease-free equilibrium remains locally asymptotically stable when the basic reproduction number is less than unity, whereas the endemic equilibrium emerges when exceeds unity. Numerical simulations further confirm that vaccination significantly reduces infection levels and accelerates recovery within the population. Graphical analysis shows that the susceptible population decreases over time while the recovered population increases steadily in the vaccinated system, as illustrated in Figs. 6 and 8. In contrast, the non-vaccinated system experiences a higher infection peak before approaching equilibrium, as observed in Fig. 10. Parameter sensitivity analysis also reveals that the transmission rate β plays a dominant role in disease spread (Fig. 12), whereas an increased recovery rate η promotes faster recovery and epidemic control (Fig. 13). The numerical comparison indicates that the Adams-Bashforth-Moulton scheme provides stable and accurate solutions compared with the fractional Euler and implicit methods, as demonstrated in Table 2 and Fig. 14, 15, 16. The convergence study confirms that the proposed numerical technique approaches second-order accuracy as the step size decreases (Table 3, Fig. 17). Overall, the fractional framework enhances the modeling capability of epidemiological systems by incorporating memory effects, improving the realism of disease dynamics. The results demonstrate that vaccination strategies and effective recovery mechanisms are essential for controlling childhood infectious diseases. Future scope Several directions may be considered to extend the present study. One possible extension is the incorporation of additional epidemiological compartments, such as exposed or quarantined individuals, which would provide a more detailed representation of disease progression. Another improvement involves integrating time-dependent vaccination rates or treatment strategies to capture dynamic public health interventions. Future research may also consider stochastic fractional models to investigate the influence of random environmental variations on disease transmission. Furthermore, the integration of optimal control strategies within the fractional modeling framework could assist policymakers in designing efficient vaccination and treatment policies. Finally, the proposed fractional modeling approach can be applied to other infectious diseases such as dengue, influenza, or COVID-19 to evaluate the effectiveness of intervention strategies under memory-dependent dynamics. Future work may also explore the integration of advanced computational techniques such as deep learning and optimal control methods, as discussed in26,27, to further enhance predictive accuracy and decision-making capabilities. Acknowledgements The author would like to express his gratitude to all reviewers for their valuable comments. Author contributions Jignesh P. Chauhan has worked on methodology and wrote the initial draft of the manuscript with all graphical and tabular results. Samaruddin Jebran has analyzed the data and helped in theoretical part. Sagar R. Khirsariya has reviewed and revised the whole manuscript. All authors approved the final version of the manuscript. Data availability All data generated or analysed during this study are clearly listed in the form of table and text in this article. Declarations Competing interests The authors declare no competing interests. Footnotes Publisher’s note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations. Jignesh P. Chauhan, Samaruddin Jebran and Sagar R. Khirsariya contributed equally to this work. Contributor Information Samaruddin Jebran, Email: jebran.samardin@ku.edu.af. Sagar R. Khirsariya, Email: ksagar108@gmail.com References Associated Data Data Availability Statement All data generated or analysed during this study are clearly listed in the form of table and text in this article.