Research article

Hardy–Littlewood maximal operators and Hausdorff operators on p-adic block spaces with variable exponents

  • In this paper, we established some sufficient conditions for the boundedness of the Hardy–Littlewood maximal operators and the Hausdorff operators on p-adic Herz spaces and p-adic local block spaces with variable exponents. In particular, the boundedness of the p-adic maximal commutator operators, the p-adic Hardy–Littlewood average operators, and the p-adic Hardy-Hilbert operators on such spaces was also discussed.

    Citation: Pham Thi Kim Thuy, Kieu Huu Dung. Hardy–Littlewood maximal operators and Hausdorff operators on p-adic block spaces with variable exponents[J]. AIMS Mathematics, 2024, 9(8): 23060-23087. doi: 10.3934/math.20241121

    Related Papers:

    [1] Sarafa A. Iyaniwura, Rabiu Musa, Jude D. Kong . A generalized distributed delay model of COVID-19: An endemic model with immunity waning. Mathematical Biosciences and Engineering, 2023, 20(3): 5379-5412. doi: 10.3934/mbe.2023249
    [2] Ugo Avila-Ponce de León, Angel G. C. Pérez, Eric Avila-Vales . Modeling the SARS-CoV-2 Omicron variant dynamics in the United States with booster dose vaccination and waning immunity. Mathematical Biosciences and Engineering, 2023, 20(6): 10909-10953. doi: 10.3934/mbe.2023484
    [3] Yuta Okada, Taishi Kayano, Asami Anzai, Tong Zhang, Hiroshi Nishiura . Protection against SARS-CoV-2 BA.4 and BA.5 subvariants via vaccination and natural infection: A modeling study. Mathematical Biosciences and Engineering, 2023, 20(2): 2530-2543. doi: 10.3934/mbe.2023118
    [4] Kimihito Ito, Chayada Piantham, Hiroshi Nishiura . Estimating relative generation times and reproduction numbers of Omicron BA.1 and BA.2 with respect to Delta variant in Denmark. Mathematical Biosciences and Engineering, 2022, 19(9): 9005-9017. doi: 10.3934/mbe.2022418
    [5] Fang Wang, Lianying Cao, Xiaoji Song . Mathematical modeling of mutated COVID-19 transmission with quarantine, isolation and vaccination. Mathematical Biosciences and Engineering, 2022, 19(8): 8035-8056. doi: 10.3934/mbe.2022376
    [6] Kento Okuwa, Hisashi Inaba, Toshikazu Kuniya . An age-structured epidemic model with boosting and waning of immune status. Mathematical Biosciences and Engineering, 2021, 18(5): 5707-5736. doi: 10.3934/mbe.2021289
    [7] Yangyang Yu, Yuan Liu, Shi Zhao, Daihai He . A simple model to estimate the transmissibility of the Beta, Delta, and Omicron variants of SARS-COV-2 in South Africa. Mathematical Biosciences and Engineering, 2022, 19(10): 10361-10373. doi: 10.3934/mbe.2022485
    [8] Kun Wang, Peng Wang, Zhengang Jiang, Lu Wang, Linhua Zhou, Dequan Qi, Weishi Yin, Pinchao Meng . Data-driven assessment of immune evasion and dynamic Zero-COVID policy on fast-spreading Omicron in Changchun. Mathematical Biosciences and Engineering, 2023, 20(12): 21692-21716. doi: 10.3934/mbe.2023960
    [9] Matthew D. Johnston, Bruce Pell, David A. Rubel . A two-strain model of infectious disease spread with asymmetric temporary immunity periods and partial cross-immunity. Mathematical Biosciences and Engineering, 2023, 20(9): 16083-16113. doi: 10.3934/mbe.2023718
    [10] Allison Fisher, Hainan Xu, Daihai He, Xueying Wang . Effects of vaccination on mitigating COVID-19 outbreaks: a conceptual modeling approach. Mathematical Biosciences and Engineering, 2023, 20(3): 4816-4837. doi: 10.3934/mbe.2023223
  • In this paper, we established some sufficient conditions for the boundedness of the Hardy–Littlewood maximal operators and the Hausdorff operators on p-adic Herz spaces and p-adic local block spaces with variable exponents. In particular, the boundedness of the p-adic maximal commutator operators, the p-adic Hardy–Littlewood average operators, and the p-adic Hardy-Hilbert operators on such spaces was also discussed.



    Since December 2019, the world has been experiencing a devastating pandemic of a novel coronavirus (COVID-19), caused by SARS-CoV-2, on a scale never before seen since the 1918/1919 influenza pandemic [1]. As of mid July 2022, the SARS-CoV-2 pandemic has caused over 555 million confirmed cases and over 6.35 million deaths globally [2,3] (with the United States bearing the brunt of the burden, with over 88.6 million confirmed cases and over 1 million COVID-19 deaths) [3]. For most parts of the year 2020, the control and mitigation efforts against SARS-CoV-2 in the United States were restricted to the use of nonpharmaceutical interventions, such as social-distancing, quarantine of suspected cases, isolation of those with symptoms of SARS-CoV-2, use face coverings, community lockdowns, contact-tracing, etc. [4,5,6,7,8], until the Food and Drug Administration (FDA) provided Emergency Use Authorization (EUA) to two safe and highly-efficacious vaccines (developed by Pfizer Inc. and Moderna Inc.) in December of 2020 [9,10]. Both of the approved vaccines were primarily administered in two-dose regiments with three to four weeks apart, and each offer an estimated protective efficacy against symptomatic COVID-19 infection of about 95% [11,12]. Another vaccine, developed by Johnson & Johnson (administered as a single dose), received FDA-EUA in late February 2021 [13] (this vaccine has an estimated 75% efficacy in preventing severe/critical illness caused by COVID-19 [14]). The rapid development and administrative deployment of effective vaccines has played an extremely vital role in minimizing and mitigating the global burden of the pandemic [15,16]. Our study is focused on these three vaccines being used and administered in the United States.

    Despite the rapid development and deployment of the effective vaccines, COVID-19 cases and mortality continued to rise in the United States for most part of 2021 (and even early 2022). This is largely due to emergence of deadly and highly-contagious SARS-CoV-2 variants of concern (notably the Alpha, Beta, Gamma, Delta and Omicron variants) [17,18,19,20,21]. Specifically, the emergence of the Omicron variant (B.1.1.529), in November of 2021, has dramatically changed the trajectory of the pandemic [22]. It is believed to be at least three times more contagious than Delta [22,23]. A subvariant of Omicron, BA.2 was first identified in the United States from a sample collected on December 14, 2021, in New Jersey [24]. It is believed to be more contagious than Omicron i.e. BA.1 [25,26].

    Numerous clinical studies have shown that the efficacy of the SARS-CoV-2 vaccines wane over time (with estimated waning time of about 9 months) [8,27,28]. Consequently, the FDA approved the administration of booster shots, for all three vaccines, during August-November of 2021. Primarily, a booster dose (for persons aged 18 years and above), were approved because of waning vaccine effectiveness over time [29]. In late March 2022, the FDA authorized a second booster shot of COVID-19 vaccines for vulnerable populations in the U.S. (i.e., for people 50 years of age and older, and for individuals with certain immuno-compromising conditions who are at higher risk of severe disease, hospitalization and death). A second booster shot is equivalent to a fourth dose for people who received a Pfizer-BioNTech or Moderna mRNA series or a third dose for those who received the single-shot Johnson & Johnson vaccine. The goal of this study is to use mathematical modeling and analysis to assess the population-level impact of vaccination and booster shots (keeping in mind the waning efficacies of all the approved vaccinations) programs, based on using the three FDA-approved vaccines, on the dynamics of the Omicron SARS-CoV-2 variant in the United States. The focus of the study is on determining the minimum vaccination coverage (i.e., vaccine-derived herd immunity) needed to effectively curtail the spread of the highly-contagious Omicron variant in the United States.

    Numerous mathematical models, of various types, have been formulated and used to gain insight and understanding on the dynamics of the COVID-19 pandemic (with majority being of the form of deterministic systems of nonlinear differential equations [4,5,7,27,30,31,32,33]). Here, too, a deterministic model will be developed and used to study the dynamics of the disease. A notable feature of the model to be developed is that it incorporates numerous pertinent aspects of the vaccination program and the current knowledge of the epidemiology of the COVID-19 pandemic, including the waning and boosting of both the vaccine-derived and natural immunity. The model will be parameterized using cumulative case data for the COVID-19 pandemic during the onset of the Omicron variant in the United States. The rest of the paper is organized as follows. The model is formulated in Section 2. The basic qualitative features of the model are also derived. Rigorous analysis of the model, with respect to the existence and asymptotic stability of its disease-free equilibrium, is carried out in Section 3. Expressions for vaccine-derived herd immunity thresholds are also derived in Section 3. The model is fitted with observed cumulative COVID-19 case data in Section 4. The methodology for implementing the data-fitting process and estimating the unknown parameters of the model is also described. Numerical simulations of the model are carried out in Section 5.

    To formulate the model for the transmission dynamics of SARS-CoV-2 in the presence of boosting and waning of both vaccine-derived and natural immunity, we split the total population at time t, denoted by N(t), into mutually exclusive compartments of unvaccinated susceptible individuals (S(t)), fully-vaccinated susceptible individuals with high vaccine-derived immunity (V1(t)), vaccinated susceptible individuals with moderate vaccine-derived immunity (V2(t)) [34], vaccinated susceptible individuals with low vaccine-derived immunity (V3(t)), exposed individuals (i.e., newly-infected individuals who are not yet infectious; E(t)), pre-symptomatic infectious individuals (Ip(t)), symptomatically-infectious individuals (Is(t)), asymptomatically-infectious individuals (Ia(t)), hospitalized individuals (Ih(t)), recovered individuals with high infection-acquired natural immunity (Rn1(t)), recovered individuals with moderate infection-acquired natural immunity (Rn2(t)), recovered individuals with low infection acquired natural immunity (Rn3(t)), recovered individuals with high infection-acquired natural and vaccine-derived immunity (Rnv1(t)), recovered individuals with moderate infection-acquired natural and vaccine-derived immunity (Rnv2(t)) and recovered individuals with low infection-acquired natural and vaccine-derived immunity (Rnv3(t)). Thus,

    N(t)=S(t)+E(t)+3i=1[Vi(t)+Rni(t)+Rnvi(t)]+j=p,s,a,hIj(t).

    Numerous clinical studies show that the vaccine-derived immunity against SARS-CoV-2 begin to wane after nine months of the receipt of the full vaccine doses [27,28,35]. Consequently, in our model formulation, individuals in the V1 class (who enjoy high level of the protective efficacy of the vaccine) are those that are within nine months of receipt of full vaccine doses. Furthermore, individuals in the V2 class are those who have received the full doses between 9 months to a year ago (hence, the vaccine efficacy is moderate). Finally, individuals in the V3 class are assumed to have received the full vaccine doses from a year to two years ago (and the vaccine efficacy is very low). This study allows for the waning and boosting of vaccine-derived and natural immunity (boosting of natural immunity is assumed to occur due to treatment or the use of other immune-boosting supplements [36,37]).

    The model is given by the following deterministic system of nonlinear differential equations, where a dot represents differentiation with respect to time t (a streamlined/abbreviated flow diagram of the model is depicted in Figure 1, and the state variables and parameters of the model are described in Tables 1 and 2, respectively):

    {˙S=Π+ωv3V3+ωn3Rn3+ωnv3Rnv3(λ+ξv+μ)S,˙V1=ξvS+ρv2V2[(1εv1)λ+ωv1+μ]V1,˙V2=ωv1V1+ρv3V3[(1εv2)λ+ωv2+ρv2+μ]V2,˙V3=ωv2V2[(1εv3)λ+ωv3+ρv3+μ]V3,˙E=λS+λ3i=1[(1εvi)Vi(t)+(1εni)Rni(t)+(1εnvi)Rnvi(t)](σE+μ)E,˙Ip=σEE(σp+γp+μ+δp)Ip,˙Is=rσpIp(ϕs+γs+μ+δs)Is,˙Ia=(1r)σpIp(γa+μ+δa)Ia,˙Ih=ϕsIs(γh+μ+δh)Ih,˙Rn1=γpIp+γsIs+γaIa+γhIh+ρn2Rn2[(1εn1)λ+ξv+ωn1+μ]Rn1,˙Rn2=ωn1Rn1+ρn3Rn3[(1εn2)λ+ξv+ωn2+ρn2+μ]Rn2,˙Rn3=ωn2Rn2[(1εn3)λ+ξv+ωn3+ρn3+μ]Rn3,˙Rnv1=ξvRn1+ρnv2Rnv2[(1εnv1)λ+ωnv1+μ]Rnv1,˙Rnv2=ξvRn2+ωnv1Rnv1+ρnv3Rnv3[(1εnv2)λ+ωnv2+ρnv2+μ]Rnv2,˙Rnv3=ξvRn3+ωnv2Rnv2[(1εnv3)λ+ωnv3+ρnv3+μ]Rnv3, (2.1)
    Table 1.  Description of the state variables of the model (2.1).
    Variable Description
    S Population of unvaccinated (wholly) susceptible individuals
    V1 Population of vaccinated susceptible individuals with high vaccine-derived immunity
    V2 Population of vaccinated susceptible individuals with moderate vaccine-derived immunity
    V3 Population of vaccinated susceptible individuals with low vaccine-derived immunity
    E Population of exposed (newly-infected individuals)
    Ip Population of pre-symptomatic infectious individuals
    Is Population of infectious individuals with clinical symptoms of the disease
    Ia Population of asymptomatically-infectious individuals
    Ih Population of hospitalized individuals
    Rn1 Population of recovered individuals with high natural immunity
    Rn2 Population of recovered individuals with moderate natural immunity
    Rn3 Population of recovered individuals with low natural immunity
    Rnv1 Population of recovered individuals with high natural and vaccine-derived immunity
    Rnv2 Population of recovered individuals with moderate natural and vaccine-derived immunity
    Rnv3 Population of recovered individuals with low natural and vaccine-derived immunity

     | Show Table
    DownLoad: CSV
    Table 2.  Description of the parameters of the model (2.1).
    Parameter Description
    Π Recruitment rate
    β Effective contact rate of individuals
    ηj(j=p,s,a,h) Modification parameter for the infectiousness of the individuals in Ip,Is,Ia, and
    Ih classes
    ξv Vaccination rate
    μ Natural death rate
    r Proportion of individuals who show clinical symptoms of the disease
    ωvi(i=1,2,3) Waning rate of vaccinated individuals in stage Vi
    ωni(i=1,2,3) Waning rate of natural immunity in individuals in stage Rni
    ωnvi(i=1,2,3) Waning rate of natural plus vaccine-derived immunity in individuals in stage Rnvi
    ρv2(ρv3) Boosting rate of vaccine-derived immunity of the individuals in stage V2(V3)
    ρn2(ρn3) Boosting rate of natural immunity of the individuals in stage Rn2(Rn3)
    ρnv2(ρnv3) Boosting rate of vaccine-derived and natural immunity of those in stage Rnv2(Rnv3)
    εvi(i=1,2,3) Average vaccine efficacy for vaccinated individuals in V1,V2 and V3 classes,
    respectively
    εni(i=1,2,3) Efficacy of natural immunity to prevent infection of recovered individuals in Rn1, \
    Rn2 and Rn3
    εnvi(i=1,2,3) Efficacy of natural and vaccine derived immunity to prevent infection of recovered
    individuals in Rnv1,Rnv2 and Rnv3
    σE Progression rate from exposed class to pre-symptomatic class
    σp Progression rate from pre-symptomatic class to either symptomatic or
    asymptomatic class
    γj(j=p,s,a,h) Recovery rate for individuals in the Ip,Is,Ia and Ih classes, respectively
    ϕs Hospitalization rate of individuals with clinical symptoms of the disease
    δj(j=p,s,a,h) Disease-induced mortality rate for individuals in the Ip,Is,Ia and Ih classes,
    respectively

     | Show Table
    DownLoad: CSV
    Figure 1.  (a) Streamlined flow diagram of the model (2.1). (b-I) - (b-IV) depict sub-flow diagrams of the model illustrating the transitions within the compartments for fully vaccinated (Vi) with high, moderate and low vaccine-derived immunity, infectious (Ij), recovered with high, moderate and low natural immunity (Rni) and recovered with both natural and vaccine-derived immunity at high, moderate and low levels (Rnvi) individuals, respectively (for i=1,2,3 and j={p,s,a,h}). The streamlined flow diagram is drawn to simplify and enhance the readability of the general structure of the model (the full version of the flow diagram of the model is given in Appendix A).

    where,

    λ=(β)(ηpIp+ηsIs+ηaIa+ηhIhN), (2.2)

    is the infection rate. In (2.2), β is the effective contact rate for individuals and ηj (with j=p,s,a,h) is the modification parameter for the heterogeneity in the infectiousness of infected individuals in the presymptomatic (Ip), symptomatic (Is), asymptomatic (Ia) and hospitalized (Ih) class, respectively.

    In the model (2.1), Π is the recruitment of individuals into the population, ωvi(i=1,2,3) is the vaccine waning rate for vaccinated individuals in stage Vi, ωni is the waning rate of natural immunity for recovered individuals in stage Rni, ωnvi is the waning rate of both vaccine-derived and natural immunity for individuals in stage Rnvi, λ is the infection rate (defined in (2.2)), ξv is the per capita vaccination rate and μ is the natural death rate. Vaccinated individuals in V2 and V3 classes receive booster doses at the rate ρvi (i=2,3) and revert to the higher efficacy vaccination stage V1 and V2, respectively. Similarly, recovered individuals in the Rn2 and Rn3 classes receive immune booster at a rate ρn2 and ρn3, respectively (and revert, respectively, to stages Rn1 and Rn2). Individuals in Rnv2 and Rnv3 (that have both the vaccine-derived and natural immunity) receive a booster at a rate ρnv2 and ρnv3, respectively (and revert to Rnv1 and Rnv2, respectively).

    The parameter εvi is the average protective efficacy of the vaccine for vaccinated susceptible individuals in the Vi(i=1,2,3) compartment, while εni(i=1,2,3) is the efficacy of natural immunity to prevent recovered individuals (in the Rni class) from acquiring future SARS-CoV-2 infection and εnvi(i=1,2,3) is the efficacy of natural and vaccine-derived immunity to prevent future SARS-CoV-2 infection of recovered individuals (in the Rnvi, i=1,2,3, classes). Exposed individuals progress to the pre-symptomatic stage at the rate σE, and pre-symptomatic individuals progress to either become symptomatically-infectious, at a rate rσp (where 0r1 is the proportion of these individuals that show clinical symptoms), or become asymptomatically-infectious, at the rate (1r)σp. Symptomatic individuals are hospitalized at a rate ϕs, and infectious individuals in stage Ij (with j=p,s,a,h) recover at a rate γj(j=p,s,a,h). Finally, disease-induced mortality occur in the Ip,Is,Ia and Ih classes at a rate δj(j=p,s,a,h).

    Some of the main assumptions made in the formulation of the model (2.1) include:

    (a) Homogeneous mixing: we assumed a well-mixed population, such that every member of the community is equally likely to mix with (and acquire infection from) every other member of the community.

    (b) Vaccinated susceptible individuals (in the V1,V2 and V3 classes) are assumed to have received the full required doses (i.e., two doses for Pfizer or Moderna vaccine, one dose for the Johnson & Johnson vaccine), and that enough time has elapsed for the body to develop immunity.

    (c) The three SARS-CoV-2 vaccines that received FDA's Emergency Use Authorization (Pfizer, Moderna and Johnson & Johnson) are imperfect [10,13,38]. That is, the vaccines offer partial protective immunity (with average efficacy 0<εvi<1), which wanes over time (at a rate ωvi), for i=1,,3 [28,35]. In other words, vaccinated individuals can experience breakthrough infection [39,40].

    (d) We assumed gradual waning of both vaccine-derived and natural immunity over time, resulting, ultimately, in reverting to the wholly-susceptible class S [28]. Moreover, the overall transitions from V1 to S, Rn1 to S and Rnv1 to S hold gamma distribution [41].

    (e) Vaccination is only offered to wholly-susceptible individuals or those who recovered naturally from COVID-19 infection but their natural immunity has waned completely or those recovered individuals who had acquired natural plus vaccine-derived immunity after recovering from COVID-19 infection but the immunity has completely waned over time. In other words, individuals who are currently infected are not vaccinated.

    (f) Immunity level can be increased or strengthened, by using immunity boosters [36,37,42,43], for the individuals in the Vi,Rni and Rnvi (i=1,,3) classes.

    Before carrying out the asymptotic analysis and numerical simulations of the model (2.1), it is instructive to explore its basic qualitative features with respect to its well-posedness (i.e., with respect to the non-negativity, boundedness and invariance of its solutions). First of all, since the model (2.1) monitors the temporal dynamics of human populations, all its parameters are non-negative. It is convenient to define the following biologically-feasible region for the model (2.1):

    Ω={(S,V1,V2,V3,E,Ip,Is,Ia,Ih,Rn1,Rn2,Rn3,Rnv1,Rnv2,Rnv3)R15+:N(t)Πμ},

    where N(t) is the total population. For the model (2.1) to be mathematically- and biologically-meaningful, it is necessary that the solutions of the model (2.1) remain non-negative for all non-negative initial conditions. That is, solutions that start in Ω remain in Ω for all time t>0 (i.e., Ω is positively-invariant with respect to the model (2.1)). Furthermore, let

    X(0)=(S(0),Vi(0),E(0),Ip(0),Is(0),Ia(0),Ih(0),Rni(0),Rnvi(0))T,

    with i=1,2,3, be the vector of initial solutions of the model (2.1). We claim the following result.

    Theorem 2.1. Consider the model (2.1) with non-negative initial data X(0). The region Ω is positively-invariant and bounded with respect to the model (2.1).

    Proof. Adding all the equations of the model (2.1) gives

    ˙N=ΠμNδpIpδsIsδaIaδhIh. (2.3)

    By the non-negativity of parameters for model (2.1), it follows from (2.3) that

    ˙NΠμN. (2.4)

    Hence, if N>Πμ, then ˙N<0. Thus, it follows, by applying a standard comparison theorem [44] on (2.4), that:

    N(t)N(0)eμt+Πμ(1eμt).

    Hence, if N(0)Πμ, then N(t)Πμ. Thus, the solutions of the model (2.1) are bounded. Therefore, every solution of the model (2.1) with initial conditions in Ω remains in Ω for all time t. In other words, the region Ω is positively-invariant and attracts all initial solutions of the model (2.1).

    The consequence of Theorem 2.1 is that it is sufficient to consider the dynamics of the flow generated by (2.1) in Ω, since the model (2.1) is epidemiologically and mathematically well-posed [45] there.

    The disease-free equilibrium (DFE) of the model (2.1) is given by:

    E0=(S,V1,V2,V3,E,Ip,Is,Ia,Ih,Rn1,Rn2,Rn3,Rnv1,Rnv2,Rnv3), (3.1)

    where,

    S=Π[μρv2A1+A2(μρv3+A3A4)]D1,V1=Πξv[ρv2A1+μρv3+A3A4]D2, V2=Πξvωv1A1D2, V3=Πξvωv1ωv2D2, (3.2)

    with,

    A1=(μ+ρv3+ωv3),A2=(μ+ωv1),A3=(μ+ωv2),A4=(μ+ωv3),B1=(μ+ξv), B2=(μ+ρv3+ωv2),B3=(μ+ρv3),B4=(μ+ρv2+ωv1),B5=(μ+ξv+ωv1), D1=μ[ξv((A2)(B2)+ρv2(A1)+ωv3(μ+ωv1+ωv2))+μρv2(A1)+(A2)(μρv3+(A3)(A4))], D2=μ(B1)[(A2)(B2)+ρv2(B3)]+μωv3[(B1)(B4)+ωv2(B5)],

    and all other components (for the infected and recovered compartments of the model) take the value zero.

    The asymptotic stability property of the DFE (E0) can be explored using the next generation operator method [46,47]. Using the notation in [46], it follows that the associated non-negative matrix of new infection terms (F) and the M-matrix of the linear transition terms (V) are given, respectively, by:

    F=[0f1f2f3f400000000000000000000]   and   V=[K10000σEK20000rσpK3000(1r)σp0K4000ϕs0K5], (3.3)

    where f1=βηp(S+AN),f2=βηs(S+AN),f3=βηa(S+AN),f4=βηh(S+AN), with

    A=(1εv1)V1+(1εv2)V2+(1εv3)V3,K1=σE+μ,K2=σp+γp+μ+δp, K3=ϕs+γs+μ+δs,K4=γa+μ+δaandK5=γh+μ+δh.

    It is convenient to define the quantity (where ρ is the spectral radius):

    Rv=ρ(FV1)={βσE(S+A)[K3K4K5ηp+K3K5ηaσp(1r)+K4K5ηsrσp+K4ηhϕsrσp](N)(5i=1Ki)}. (3.4)

    The result below follows from Theorem 2 of [46].

    Theorem 3.1. The disease-free equilibrium (E0) of the model (2.1) is locally-asymptotically stable (LAS) if Rv<1, and unstable if Rv>1.

    The threshold quantity Rv is the vaccination reproduction number of the model (2.1), which measures the average number of new COVID-19 cases generated by a single infectious individual introduced into a population where a certain proportion is vaccinated. The epidemiological interpretation of Theorem 3.1 is that a small influx of COVID-19 cases will not generate a large outbreak in the community if the vaccination reproduction number (Rv) is brought to, and maintained at a, value less than unity.

    In the absence of vaccination and other public health interventions, the vaccination reproduction number (Rv) reduces to the basic reproduction number (denoted by R0). That is,

    R0=Rv|V1=V2=V3=0={βσE[K3K4K5ηp+K3K5ηaσp(1r)+K4K5ηsrσp+K4ηhϕsrσp]5i=1Ki}.

    Consider the special case of the model (2.1) with perfect vaccine protective efficacy against primary infection and re-infection and no waning of vaccine-derived and natural immunity (i.e., we consider the model (2.1) with εvi=εni=εnvi=1, ωni=ωnvi=0, with i=1,2,3). It is convenient to let:

    ˜S=S+V1+V2+V3. (3.5)

    Substituting εvi=εni=εnvi=1 and ωni=ωnvi=0 (with i=1,2,3) into the model (2.1), it follows that the equation for the rate of change of the new compartment ˜S (defined in (3.5)) is given by (where the infection rate, λ, is as defined in (2.2)):

    d˜Sdt=Πμ˜SλS. (3.6)

    It can be shown that the special case of the model is positively-invariant and bounded in the region (as shown in Section 2.1)

    Ω={(˜S,E,Ip,Is,Ia,Ih,Rn1,Rn2,Rn3,Rnv1,Rnv2,Rnv3)R12+:N(t)Πμ}.

    Furthermore, the disease-free equilibrium of the special case of the model is given by:

    E0R=(˜S,E,Ip,Is,Ia,Ih,Rn1,Rn2,Rn3,Rnv1,Rnv2,Rnv3)=(Πμ,0,0,0,0,0,0,0,0,0,0,0). (3.7)

    For the aforementioned special case of the model, it can be seen that the associated next generation matrix of new infection terms, denoted by ˜F, is given by (note that, for this special case, the next generation matrix of linear transition terms, V, remains the same, as given in (3.3). Further, N=Π/μ):

    ˜F=[0βηp(˜SN)βηs(˜SN)βηa(˜SN)βηh(˜SN)00000000000000000000].

    Thus,

    ˜Rv=ρ(˜FV1)={βσE˜S[K3K4K5ηp+K3K5ηaσp(1r)+K4K5ηsrσp+K4ηhϕsrσp](N)(5i=1Ki)}. (3.8)

    We claim the following result:

    Theorem 3.2. Consider the special case of the model (2.1) with εvi=εni=εnvi=1 and ωni=ωnvi=0 (for i=1,2,3). The disease-free equilibrium of the special case of the model (E0R) is globally-asymptotically stable in Ω whenever ˜Rv<1.

    Proof. Consider the model (2.1) with εvi=εni=εnvi=1 and ωni=ωnvi=0 (i=1,2,3). Further, let ˜Rv<1. The proof is based on using a comparison theorem [48]. It can be shown, first of all, that the region Ω is positively-invariant and attracts all solutions of the aforementioned special case of the model (2.1) (as shown in Section 2.1). The equations for the infected compartments of the special case of the model (2.1) can be re-written in terms of the next generation matrices (˜F and V) as below:

    ddt[E(t)Ip(t)Is(t)Ia(t)Ih(t)]=(˜FV)[E(t)Ip(t)Is(t)Ia(t)Ih(t)]M[E(t)Ip(t)Is(t)Ia(t)Ih(t)], (3.9)

    where,

    ˜FV=[(σE+μ)βηp(˜SN)βηs(˜SN)βηa(˜SN)βηh(˜SN)σE(σp+γp+μ+δp)0000rσp(ϕs+γs+μ+δs)000σp(1r)0(γa+μ+δa)000ϕs0(γh+μ+δh)],

    and,

    M=β(1SN)[0ηpηsηaηh00000000000000000000]. (3.10)

    Since S(t)N(t) for all t>0 in Ω, it follows that the matrix M, defined in (3.10), is non-negative. Hence, Eq (3.9) can be re-written in terms of the following inequality:

    ddt[E(t)Ip(t)Is(t)Ia(t)Ih(t)](˜FV)[E(t)Ip(t)Is(t)Ia(t)Ih(t)]. (3.11)

    It should be recalled from the local asymptotic stable result for the DFE (given in Theorem 3.1) that all eigenvalues of the next generation matrix ˜FV1 are negative if ˜Rv<1 (i.e., ˜FV is a stable matrix). Thus, it can be concluded that the linearized differential inequality system (3.11) is stable whenever ρ(˜FV1)<1. Hence, it follows that (for the linear system of ordinary differential equations (3.11)):

    (E(t),Ip(t),Is(t),Ia(t),Ih(t))(0,0,0,0,0),ast.

    Substituting E(t)=Ip(t)=Is(t)=Ia(t)=Ih(t)=0 into the differential equations for the rate of change of the Rni(t), Rnvi(t) and ˜S (with i=1,2,3) compartments of the model (2.1) shows that (where ˜S is defined in (3.7)):

    Rni(t)0,Rnvi(t)0(withi=1,,3)and˜S(t)˜S,ast.

    Thus, the DFE (E0R) of the special case of the model (2.1), with εvi=εni=εnvi=1 and ωni=ωnvi=0 (i=1,2,3), is globally-asymptotically stable in Ω whenever ˜Rv<1.

    Epidemiologically-speaking, Theorem 3.2 shows that, for the special case of the model (2.1) with εvi=εni=εnvi=1 and ωni=ωnvi=0 (i=1,2,3), the disease can be eliminated from the community if the threshold quantity, ˜Rv, can be brought to (and maintained at) a value less than unity.

    The global asymptotic stability of the disease-free equilibrium of the model (2.1) can also be established for another special case, as described below. Consider the special case of the model (2.1) in the absence of disease-induced mortality (i.e. δp=δs=δa=δh=0) and no reinfection (i.e., εni=εnvi=1 with i=1,2,3). Setting δp=δs=δa=δh=0 in the model (2.1), and adding all the equations of the model shows that dNdt=ΠμN, from which it follows that N(t)Πμast. From now on, we replace N(t) with its limiting value, N=Π/μ (i.e., the standard incidence formulation for the infection rate is now replaced by a mass action incidence). Furthermore, it is convenient to define the following feasible region for the special case of the model (where S and Vi, with i=1,2,3, are as defined in Section 3):

    Ω={(S,V1,V2,V3,E,Ip,Is,Ia,Ih,Rn1,Rn2,Rn3,Rnv1,Rnv2,Rnv3)Ω:SS,ViVi,i=1,2,3}.

    It can be shown that the region Ω is positively-invariant with respect to the aforementioned special case of the model [49]. Further, for this special case of the model, it can be shown that the associated next generation matrices are given, respectively, by:

    ˆF=[0^f1^f2^f3^f400000000000000000000]   and   ˆV=[^K10000σE^K20000rσp^K3000(1r)σp0^K4000ϕs0^K5],

    where (with S and A as defined in Section 3),

    ^f1=βηp(S+A),^f2=βηs(S+A),^f3=βηa(S+A),^f4=βηh(S+A),

    and,

    ^K1=σE+μ,^K2=σp+γp+μ,^K3=ϕs+γs+μ,^K4=γa+μand^K5=γh+μ.

    It is convenient to define the following threshold quantity:

    ˆRv=ρ(ˆF^V1)={βσE(S+A)[^K3^K4^K5ηp+^K3^K5ηaσp(1r)+^K4^K5ηsrσp+^K4ηhϕsrσp](5i=1ˆKi)}. (3.12)

    We claim the following result:

    Theorem 3.3. Consider the special case of the model (2.1) in the absence of disease-induced mortality (i.e., δp=δs=δa=δh=0) and no reinfection of recovered individuals (i.e., εni=εnvi=1 with i=1,2,3). The disease-free equilibrium of the special case of the model (E0) is globally-asymptotically stable in Ω whenever ˆRv<1.

    The proof of Theorem 3.3, based on using a comparison theorem, is given in Appendix B.

    Herd immunity, which is a measure of the minimum percentage of the number of susceptible individuals that need to be protected against the infection in order to eliminate community transmission of an infectious disease, can be attained through two main ways, namely natural immunity route (following natural recovery from infection with the disease) or by vaccination (which is widely considered to be the safest and the fastest way) [50,51]. For vaccine-preventable diseases, such as COVID-19, it is not practically possible to vaccinate every susceptible individual in the community due to various reasons, such as individuals with certain underlying medical conditions, infants, individuals who are pregnant, breastfeeding women or those who are unwilling to be vaccinated for COVID-19 due to some other reasons [30]. Therefore, it is crucial to determine the minimum proportion of the susceptible population that need to be vaccinated in order to protect those that cannot be vaccinated (so that vaccine-induced herd immunity is achieved in the population). Since we have three vaccination classes (V1, V2 and V3), accounting for the three levels of vaccine-derived immunity (high, moderate and low), we will compute vaccine-derived herd immunity thresholds for the United States with respect to each of the vaccination classes. Specifically, we let

    fv1=V1N,fv2=V2N,fv3=V3N,withN=Πμ. (3.13)

    Here, fv1 represents the proportion of susceptible members of the population that have been fully-vaccinated (using any of the three approved vaccines), fv2 represents proportion of V2 boosted and fv3 represents proportion of V3 boosted, at the disease-free equilibrium (E0). Using the definition (3.13) in (3.4) gives:

    Rv={βσE[1(εv1fv1+εv2fv2+εv3fv3)][K3K4K5ηp+K3K5ηaσp(1r)+K4K5ηsrσp+K4ηhϕrσp](5i=1Ki)},

    which can be expressed in terms of the basic reproduction number (R0) as:

    Rv=(13i=1εvifvi)R0. (3.14)

    Setting Rv=1 (i.e., the bifurcation point) in Eq (3.14), and simplifying, gives:

    3i=1εvifvi=(11R0), (3.15)

    from which we can solve for the fraction fully-vaccinated (those in V1 class), or received booster doses (i.e., those in V2 and V3 classes), at steady-state for each vaccinated class (denoted by fviwithi=1,,3), in terms of the basic reproduction number, giving:

    fv1=1εv1[(11R0)(εv2fv2+εv3fv3)]=fcv1(forR0>1), fv2=1εv2[(11R0)(εv1fv1+εv3fv3)]=fcv2(forR0>1), fv3=1εv3[(11R0)(εv1fv1+εv2fv2)]=fcv3(forR0>1). (3.16)

    It follows from Eq (3.16) that Rv<(>)1 if fvi>(<)fcvi (withi=1,2,3). Furthermore, Rv=1 whenever fvi=fcvi (withi=1,2,3). This result is summarized below:

    Theorem 3.4. Vaccine-induced herd immunity (i.e., COVID-19 elimination) can be achieved in the United States, using any of the approved anti-COVID vaccine, if vaccination of susceptible individuals and boosting of vaccine-derived immunity resulted in fvi>fcvi (i.e., if Rv<1) for each corresponding i, for all i=1,2,3. If fvi<fcvi (i.e., if Rv>1), then the vaccination program will fail to eliminate the COVID-19 pandemic.

    Epidemiologically-speaking, Theorem 3.4 implies that the use of any of the approved COVID-19 vaccines can lead to the elimination of the pandemic in the United States if the proportion of susceptible individuals fully-vaccinated and with high level of vaccine-derived immunity (i.e., those in V1 class) and boosted (i.e., those in V2 and V3 classes) at steady-state reached or exceeded the aforementioned critical threshold values. In other words, the SARS-CoV-2 pandemic is predicted to be eliminated in the United States if fvi>fcvi for each corresponding i, for all i=1,2,3. On the contrary, the Vaccination program will fail to eliminate the pandemic if the proportion vaccinated (and boosted) at the disease-free equilibrium falls below the aforementioned critical herd immunity thresholds.

    It should be mentioned that since the Pfizer and Moderna vaccines offer protective efficacy of about 95% and 94%, respectively [52], and the Johnson & Johnson vaccine offers a protective efficacy of about 67% [53], we set the average vaccine protective efficacy for individuals in the V1 class to be

    εv1=0.95+0.94+0.6730.85. (3.17)

    It is worth mentioning that the expression (3.17), for the average vaccine protective efficacy, is a dynamic quantity that depends on the actual combinations of SARS-CoV-2 vaccines adopted in the community at time t (it should be mentioned that, in deriving the estimate for εv1 in Eq (3.17), we used the values of the respective efficacies of the three vaccines estimated in [52,53] during the period around October, 2020 to January, 2021). Table 5 summarizes the assumed baseline efficacy levels for average vaccine-derived and natural immunity to be used in our numerical simulations. Using the baseline values of the fixed and fitted parameters in Tables 3 and 4, together with the baseline average vaccine-derived and natural immunity protective efficacy levels in Table 5, it follows from Eq (3.16) that the critical vaccine-derived herd immunity threshold for each of the vaccinated compartment is given, respectively, by fcv1=0.59, fcv2=0.42 and fcv3=1.01. In other words, based on the parameterization of the model (2.1) with the recent case data for Omicron BA.1 variant in the United States, population-level herd immunity can be achieved in the United States if the following conditions hold:

    Table 3.  Baseline values of the fixed parameters of the model (2.1).
    Parameter Value Source Parameter Value Source
    σE 1/3 day1 [60] ωv1 1/274 day1 [61]
    σp 1/2 day1 [62] ωv2 1/365 day1 [61]
    r 0.099 (dimensionless) [63] ωv3 1/365 day1 [61]
    ϕs 1/5 day1 [27] ωn1 1/274 day1 [61]
    γs 1/10 day1 [28,53] ωn2 1/365 day1 [61]
    γa 1/5 day1 [28,53] ωn3 1/365 day1 [61]
    γh 1/8 day1 [28,53] ωnv1 1/548 day1 Assumed
    ηp 5/4 (dimensionless) [28] ωnv2 1/730 day1 Assumed
    ηs 1/2 (dimensionless) [8] ωnv3 1/730 day1 Assumed
    ηa 3/2 (dimensionless) [28] ρv2 1/14 day1 [56,64,65]
    ηh 3/20 (dimensionless) [8] ρv3 1/14 day1 [56,64,65]
    ξv 1.9 × 105 day1 [66] ρn2 1/14 day1 [64]
    Π 11400 day1 [8] ρn3 1/14 day1 [64]
    μ 3.4 × 105 day1 [8] δh 5.0 × 105 day1 [8]
    δp 0 day1 [67] δa 0 day1 [68]

     | Show Table
    DownLoad: CSV
    Table 4.  Baseline values of fitted (estimated) parameters of the model (2.1), obtained by fitting the model with the observed cumulative daily COVID-19 data for the United States for the period November 28th, 2021 to February 23rd, 2022.
    Parameter Estimated Value Parameter Estimated Value
    β 0.2120 day1 ρnv2 0.1996 day1
    ρnv3 0.6398 day1 δs 4.9804 × 105 day1

     | Show Table
    DownLoad: CSV
    Table 5.  Assumed baseline levels of the parameters for the efficacy of the average vaccine-derived and natural immunity.
    Vaccine Efficacy of Vn class Vaccine Efficacy of Rn class Vaccine Efficacy of Rnv class
    εv1=0.85 εn1=0.85 εnv1=0.95
    εv2=0.50 εn2=0.50 εnv2=0.50
    εv3=0.20 εn3=0.20 εnv3=0.20

     | Show Table
    DownLoad: CSV

    (a) at least 59% of the wholly-susceptible individuals are fully vaccinated (i.e., 59% of individuals in the S class are fully-vaccinated and moved to the V1 class);

    (b) 42% of vaccinated individuals with moderate vaccine-derived immunity (i.e., those in V2 class) are boosted;

    (c) almost all of the vaccinated individuals whose level of vaccine-derived immunity is low (i.e., those in V3) are boosted.

    Hence, vaccine-derived herd immunity will be achieved via the aforementioned vaccination-boosting strategy that entails having at least 59% of the wholly-susceptible population to be fully-vaccinated followed by the boosting of an average of (42%+101%)/2 = 71.5% of the fully-vaccinated individuals with moderate and low vaccine-derived immunity.

    Figure 2 depicts contour plots of the vaccination reproduction number (Rv), as a function of vaccination efficacy (εvi) and coverage of fully-vaccinated or boosted individuals at steady-state (fvi), for i=1,2,3. It follows from these plots that, for the overall average vaccine-protective efficacy set at 85% (as stated above), at least 59% of the wholly-susceptible population need to be vaccinated at steady-state to bring the vaccination reproduction number (Rv) below one (Figure 2(a)). For the case when the average vaccine-protective efficacy has waned to 50% (i.e., fully-vaccinated individuals now have moderate vaccine-derived immunity; here, εv2=0.5, as given in Table 5), up to 42% of individuals in the V2 class need to be boosted to bring the vaccination reproduction number to a value below one (Figure 2(b)). Finally, when vaccine-derived immunity has waned to the low level of 20%, the contour plot in Figure 2(c) shows that all of the fully-vaccinated individuals with low vaccine-derived immunity (i.e., individuals in the V3 class) need to be boosted to bring the reproduction number to a value less than one. In summary, the results depicted in Figure 2 show that population-level herd immunity can be achieved in the United States via the implementation a vaccination program (based on using any of the three approved vaccines) that emphasizes the full vaccination of a sizable proportion of the susceptible pool (at least 59%) followed by the administration of booster doses to individuals in whom their vaccine-derived immunity has waned to moderate (at least 42%) or low (at least 100%) levels. Overall, our study shows that, for the case where the protective immunity offered by the vaccine for fully-vaccinated individuals in the V1 class is 85%, vaccine-derived herd immunity can be achieved in the United States if at least 59% of the susceptible population is fully-vaccinated (with any of the three approved vaccines) followed by the boosting of at least 71.5% of the fully-vaccinated individuals whose vaccine-derived immunity has waned to moderate or low level.

    Figure 2.  Contour plots of the vaccine reproduction number (Rv) of the model (2.1), as a function of vaccine coverage or boosting at steady- state (fvi; i=1,2,3) and average vaccine efficacy (εvi; i=1,2,3), for the United States. (a) Vaccination of wholly-susceptible individuals (S(t); fv1 is proportion of wholly-susceptible individuals who are fully-vaccinated at steady-state). (b) Boosting of vaccinated individuals with moderate vaccine-derived immunity (V2(t); fv2 is the proportion of vaccinated individuals in the V2 class who are boosted). (c) Boosting of vaccinated individuals with low vaccine-derived immunity (V3(t); fv3 is the proportion of vaccinated individuals in the V3 class who are boosted). Parameter values used in these simulations are as given by their respective baseline values in Tables 35.

    It should be mentioned that for the case when the high level of the average vaccine-induced efficacy for individuals in the V1 class is decreased to 55%, for instance (while the average vaccine protective efficacy for individuals in the V2 and V3 classes remain at the baseline level), our simulations showed that at least 91% of the wholly-susceptible population need to be fully-vaccinated, followed by the (marginal) boosting (about 4%) of the vaccinated individuals whose vaccine-derived immunity has waned to moderate or low level. Thus, this study shows that lower average protective efficacy of the vaccines (for fully-vaccinated individuals) incurs higher requirement for the vaccination coverage of the susceptible population (followed by a correspondingly low level of boosting for the vaccinated individuals whose average vaccine protective efficacy has waned to moderate or low level) to achieve herd immunity. Vaccinating 91% of the wholly-susceptible population is, of course, not realistically feasible in large populations, such as the United States. Hence, it is imperative that highly efficacious vaccines are developed and used, and combined with boosting (at moderate to high levels) of vaccinated individuals whose immunity has waned to moderate or low level. In other words, using vaccines with higher protective efficacy (e.g., vaccines with 85% average protective efficacy, as computed in Eq (3.17)) incurs lower, and realistically attainable, requirement for the vaccination coverage (about 59%) and attainable (moderate to high) boosting level (about 71.5%) for the fully-vaccinated individuals whose vaccine-derived immunity has waned to moderate or low level. As of May 20, 2022, data from the CDC shows that about 66.5% of the U.S. population is fully-vaccinated, and about 46.4% of the population of fully-vaccinated individuals is boosted [54]. Thus, this study shows that, even for the scenario where the three vaccines offer such high average protective efficacy, a sizable proportion of the fully-vaccinated individuals need to be boosted in order to achieve vaccine-derived herd immunity in the United States (using the aforementioned combined vaccination-boosting strategy).

    In this section, we fit the model (2.1) by using the available data for the observed cumulative COVID-19 cases for the U.S. (for the period November 28, 2021–February 23, 2022). The model (2.1) has several parameters, some of which are known from the literature (as tabulated in Table 3) and the remaining unknown parameters are obtained by fitting the model (2.1) with the cumulative case data obtained from the Johns Hopkins University COVID-19 repository [3]. The model was fitted using a standard nonlinear least squares approach, which involved using the inbuilt MATLAB minimization function "lsqcurvefit" to minimize the sum of the squared differences between each observed cumulative cases data points and the corresponding cumulative cases points obtained from the model (2.1) (i.e., rσpIp). The unknown parameters which are estimated from the fitting are presented in Table 4.

    The data fitting is done by splitting the available COVID-19 cumulative case data for the United States for the period from November 28, 2021 (when Omicron first emerged) to March 23, 2022 into two segments. The first segment of the data, from November 28, 2021 to February 23, 2022 (i.e., the region to the left of the dashed vertical cyan line), was used to fit the model (2.1) and to estimate the unknown parameters. The results obtained, depicted in Figure 3(a), show a very good fit for the model output (blue curve) and the observed cumulative case data (red dots). The data for the second segment of the data, for the period from February 24th, 2022 to March 23, 2022, was used for the cross validation of the fitted data. This also shows a very good fit for the model output (green curve) and for the remaining data points of the observed data (red dots) of Figure 3(a). This segment of the Figure 3(a) clearly shows that the model (2.1) cross validates the observed cumulative case data for the period from February 23, 2022 to March 23, 2022 perfectly (solid green curve). Furthermore, we show, in this figure, the prediction of the model for the cumulative COVID-19 cases for approximately a five-week period after March 24, 2022 (i.e., the region to the right of the dashed vertical black line), as illustrated by the solid magenta curve in Figure 3(a). The model was then simulated (using the fixed, fitted and assumed parameter baseline values in Tables 3, 4 and 5), and compared with the observed new daily cases data. The results obtained, depicted in Figure 3(b), show a very good fit for the daily new COVID-19 cases in the United States.

    Figure 3.  (a) Time series illustration of the least squares fit of the model (2.1), showing the model's output for the cumulative daily cases in the United States (blue curve) compared to the observed cumulative confirmed cases for the United States (red dots) from November 28, 2021 to February 23, 2022 (segment to the left of the dashed vertical cyan line). (b) Simulation result of (2.1), showing observed daily confirmed COVID-19 cases for the United States as a function of time, using the fixed, estimated and assumed baseline parameter values given in Tables 3, 4 and 5 respectively. The segment from February 24, 2022 to April 30, 2022 (i.e., solid green and magenta curves or the entire segment to the right of the dashed cyan vertical line) illustrates the performance of the model (2.1) in predicting the cumulative and daily cases in the United States.

    Furthermore, in fitting the model to observed cumulative case data for the United States, we set the average efficacies for individuals in the Vi classes (with i=1,2,3) to their respective values in Table 5. Specifically, the value of the average vaccine efficacy for individuals in the V1 class (εv1) is set at 85% based on the expression given in Eq (3.17). The estimate for the average vaccine efficacy for individuals in the V2 class (εv2) was adapted from the empirical studies in [55,56]. First of all, these studies show variations in the timing of the attainment of the residual efficacy (after a few months of receipt of COVID-19 vaccination), which depend on the type of vaccine (Pfizer, Moderna or Johnson & Johnson). Further, the studies showed that the vaccine-derived efficacy of the Johnson & Johnson vaccine dropped from the initial 74.8% to 59.4% after five months of the receipt of the single-dose [55]. Similarly, the protective efficacy of the Pfizer vaccine decreased from 94.5% to 75.7% after 8 months of the receipt of the full doses of the Pfizer vaccine [56]. Finally, the efficacy of the Moderna vaccine decreased from 95.5% to 84.3% after 5 months of full vaccination [56]. Based on these estimates, we considered it plausible to set the average vaccine efficacy for individuals in the V2 class to be 50%.

    Additionally, numerous empirical studies have shown that the effectiveness of the Pfizer and Moderna vaccines against symptomatic COVID-19 is less than 20% for the Omicron variant after the administration of the second dose of the vaccine [57,58]. Consequently, we deem it reasonable to set the average vaccine-derived efficacy for individuals in the V3 class to be about 20%. Moreover, an empirical study showed that individuals who received the second booster dose of the Pfizer or Moderna vaccine have their vaccine effectiveness against the acquisition of breakthrough infection increased from 19% to 49% [59]. This is in line with the assumption in our model formulation regarding the transition of individuals from the V3 class (where the average efficacy was set to be 20%) to the V2 class (where the average vaccine efficacy is set at 50%).

    The model (2.1) will now be simulated to assess the population-level impact of waning and boosting of vaccine-derived and natural immunity on the dynamics of the Omicron variant in the United States. Unless otherwise stated, the simulations will be carried out using the baseline values of the parameters tabulated in Tables 35.

    To assess the impact of waning of vaccine-derived immunity for this scenario, we simulate the model (2.1) using the following three (arbitrarily-chosen) levels of the parameters related to the waning of vaccine-derived immunity in the population:

    (i) Low level of waning of vaccine-derived immunity: here, we consider vaccine-derived immunity to wane within 48 months (i.e., we set ωv1=ωv2=ωv3=0.0007 per day) but parameters related to natural immunity and combined natural and vaccine-derived immunity are kept at their respective baseline levels.

    (ii) Baseline level of waning of vaccine-derived immunity: in this case, waning of vaccine-derived immunity is set to occur within 9 months (so that, ωv1=ωv2=ωv3=0.0037 per day) but parameters related to natural immunity and combined natural and vaccine-derived immunity are kept at their respective baseline levels.

    (iii) High level of waning of vaccine-derived immunity: in this scenario, vaccine-derived immunity wanes within 3 months (i.e., ωv1=ωv2=ωv3=0.0110 per day), but parameters related to natural immunity and combined natural and vaccine-derived immunity are kept at their respective baseline levels.

    For these simulations, all other parameters of the model (including those that involve the waning of natural immunity and combined waning of natural and vaccine-derived immunity) are maintained at their baseline values (given in Tables 35). Furthermore, these simulations are carried in the absence and presence of boosting of vaccine-derived immunity (recall that boosting of vaccine-derived immunity, maintained at baseline level, is achieved via the administration of the required doses of any of the approved SARS-CoV-2 booster vaccines used in the United States).

    The simulation results obtained, depicted in Figure 4. First of all, these simulations also depict the fitting of the model's output for the daily new and cumulative cases with the observed data (used in Section 4) for the baseline scenario (as shown by the blue curves and the red dots in Figure 4). Furthermore, these simulations show that, in the absence of boosting of vaccine-derived immunity, waning of vaccine-derived immunity generally induces only a marginal impact on the average number of new daily COVID-19 cases in the United States, for each of the three waning levels considered in our simulations, in comparison to the baseline scenario. For example, under the fast waning scenario for vaccine-derived immunity (i.e., vaccine-derived immunity wanes within three months, but natural immunity is maintained at its baseline level) and no boosting of vaccine-derived immunity is implemented, the simulations show a marginal (about 2%) increase in the peak level of the daily new cases, in comparison to the peak baseline level (this is evident by comparing the blue and magenta curves in Figure 4(a), and the zoomed-in version of the segments of the curves near the peaks shown in Figure 4(b)). For the slow waning scenario (i.e., if the vaccine-derived immunity wanes within 48 months, but natural immunity is maintained at baseline level), the increase in daily new cases at the peak (in comparison to the baseline) reduces to about 1.5% (compare the blue and green curves in Figures 4(a) and (b)).

    Figure 4.  Simulations of the model (2.1) to assess the population-level impact of waning of vaccine-derived immunity in the absence and presence of boosting of vaccine-derived immunity (maintained at baseline level). (a)(d): average number of new daily cases at the peak in the absence ((a) and (b)) and presence ((c) and (d)) of boosting of vaccine-derived immunity. (e)(h): cumulative number of new cases in the absence ((e) and (f)) and presence ((g) and (h)) of boosting of vaccine-derived immunity (maintained at baseline level). Three levels of waning of vaccine-derived immunity were considered: vaccine-derived immunity wanes in three months (magenta curves), nine months (blue curves) and forty eight months (green curves). Zoomed-in versions of the portions of the curves near the peaks depicted in Figures (a) and (c) are shown in Figures (b) and (d), respectively. Similarly, zoomed-in versions of the portions of the curves near the peaks in Figures (e) and (g) are shown in Figures (f) and (h), respectively. The values of the other parameters of the model used in these simulations are as given in Tables 35.

    In the presence of boosting of vaccine-derived immunity (at baseline level), our simulations show a significant reduction in the average number of daily new cases at the peak recorded under the above waning scenarios without boosting of vaccine-derived immunity. For instance, for the case where vaccine-derived immunity wanes within three months (but natural immunity is maintained at baseline level), boosting of vaccine-derived immunity at the baseline level significantly reduces the increase in daily new cases at the peak (by about 90%), in comparison to the corresponding case without boosting of vaccine-derived immunity (compare the blue and magenta curves in Figure 4(c), and the corresponding zoomed-in portions of the curves near the peaks shown in Figure 4(d)). Furthermore, under the slow waning scenario, boosting of vaccine-derived immunity at baseline level further increases the reduction in the peak daily new cases (compare the green and blue curves in Figures 4(c) and (d)). Similar dynamics are observed (and illustrated) with respect to the cumulative number of new cases, for the three waning scenarios considered in these simulations without (Figures 4(e) and (f)) and with (Figures 4(g) and (h)) boosting of vaccine-derived immunity.

    We further simulated the model to assess the impact of waning and boosting of vaccine-derived immunity (for the case where natural immunity is maintained at baseline) for the following two scenarios:

    Scenario (a): Fast waning and slow boosting. Here, we assume that the waning of vaccine-derived immunity range between 3 to 6 months and the duration of boosting of vaccine-derived immunity range from 20 days to 180 days.

    Scenario (b): Fast waning and boosting near the baseline level. Under this scenario, vaccine-derived immunity wanes within the same 3 to 6 months period (as in Scenario (a)), but boosting of vaccine-derived immunity is accelerated to be implemented within 10 to 20 days (i.e., near the baseline level of 14 days).

    The results obtained are depicted in the form of heat maps for the vaccination reproduction number (Rv) of the model (2.1), as a function of the rates of waning (ωv) and boosting (ρv) of vaccine-derived immunity in Figure 5. This figure shows that, for the fast waning and slow boosting scenario (i.e., Scenario (a)), the values of the vaccination reproduction number lie in the range Rv[0.82,1.21] (with a mean of Rv1.015), suggesting that the disease will persist in the population (this is in line with the theoretical result given in Theorem 3.1). In other words, this result shows that faster waning and slower boosting, in comparison to waning and boosting at baseline levels, increases the prospect for disease persistence in the population. For Scenario (b), our simulations (Figure 5(b)) show a marked decrease in the range of the reproduction number, with R[0.78,0.89] (with a mean of Rv=0.835), suggesting possible elimination of the pandemic (in line with Theorems 3.1–3.3). Thus, boosting of vaccine-derived (near the baseline rate) enhances the prospect for pandemic elimination.

    Figure 5.  Effect of waning and boosting of vaccine-derived immunity. Heat maps of the vaccination reproduction number (Rv), as a function of the rates of waning (ωv) and boosting (ρv) of vaccine-derived immunity. (a) Waning of vaccine-derived immunity range between 3 to 6 months, and duration of boosting of vaccine-derived immunity range from 20 days to 180 days (slow boosting). (b) Waning of vaccine-derived immunity range from 3 to 6 months, while duration of boosting of vaccine-derived immunity range from 10 to 20 days (fast boosting).

    In summary, while the simulations in this section show that waning of vaccine-derived immunity generally induces only a marginal impact in the average number of new cases at the peak of the COVID-19 pandemic, boosting of vaccine-derived immunity (maintained at its baseline level) resulted in a dramatic reduction in the average number of new cases at the peak, in comparison to the case where boosting is not implemented. Furthermore, delay in boosting of vaccine-derived immunity, in comparison to the baseline level of boosting, could alter the trajectory of the disease from possible elimination (as measured by the vaccine reproduction number, Rv, taking a value less than one) to persistence of the disease (as measured by the reproduction number being greater than one).

    Natural immunity can be boosted via treatment or the use of other immune-boosting supplements [36,37]. To assess the impact of waning of natural immunity, we simulated the model (2.1) using the following (arbitrarily-chosen) waning levels:

    (i) Low level of waning of natural immunity: here, too, we consider natural immunity to wane within 48 months (i.e., we set ωn1=ωn2=ωn3=0.0007 per day), but vaccine-derived immunity and combined natural and vaccine-derived immunity are kept at their respective baseline levels.

    (ii) Baseline waning of natural immunity: in this case, waning of natural immunity is set to occur within 9 months (so that, ωn1=ωn2=ωn3=0.0037 per day), but vaccine-derived immunity and combined natural and vaccine-derived immunity are maintained at their respective baseline levels.

    (iii) High level of waning of natural immunity: here, too, natural immunity is assumed to wane within 3 months (i.e., ωn1=ωn2=ωn3=0.0110 per day), but vaccine-derived immunity and combined natural and vaccine-derived immunity are kept at their respective baseline levels.

    For the simulations in this section, we set all other parameters (including those related to the waning of vaccine-derived immunity and combined waning of natural and vaccine-derived immunity) to their baseline values (given in Tables 35). The simulation results obtained, depicted in Figure 6, also showed that waning of the natural immunity in general only induces a marginal increase in the average number of new daily COVID-19 cases in the United States, in comparison to the baseline scenario (where the waning of natural immunity is assumed to occur within 9 months). In particular, if natural immunity wanes within three months and no boosting of natural immunity is implemented, the average number of new daily cases at the peak increases by about 4.2%, in comparison to the baseline scenario (compare the blue and magenta curves in Figure 6(a), and the zoomed-in portions of the curves near the peaks, depicted in Figure 6(b)). An additional marginal increase in the average number of new daily cases at the peak is recorded under the slow waning scenario for the natural immunity, in comparison to the baseline scenario (compare the blue and green curves in Figures 6(a) and (b)).

    Figure 6.  Simulations of the model (2.1) to assess the population-level impact of waning of natural immunity for the case with and without boosting of natural immunity (at the baseline level). (a)(d): average number of new daily cases at the peak in the absence ((a) and (b)) presence ((c) and (d)) of boosting of natural immunity. (e)(h): cumulative number of new cases in the absence ((e) and (f)) and presence ((g) and (h)) of boosting of natural immunity (maintained at baseline level). Three levels of waning of natural immunity were considered: natural immunity wanes in three months (magenta curves), nine months (blue curves) and forty eight months (green curves). Zoomed-in versions of the portions of the curves near the peaks depicted in Figures (a) and (c) are shown in Figures (b) and (d), respectively. Similarly, zoomed-in versions of the portions of the curves near the peaks in Figures (e) and (g) are shown in Figures (f) and (h), respectively. The values of the other parameters of the model used in these simulations are as given in Tables 35.

    However, if natural immunity is boosted (at baseline level), our simulations show a marked reduction in the increase in the average daily new cases recorded at the peak, in comparison to the corresponding scenario without boosting of the natural immunity. Specifically, when natural immunity wanes within three months and boosting of natural immunity is implemented (maintained at its baseline level), the increase in the average number of new daily cases at the peak (in comparison to the baseline) reduces to about 1.6% (compare the blue and magenta curves in Figure 6(c), and the zoomed-in portions near the peak depicted in Figure 6(d)). This represents an approximately 62.2% reduction in the average daily new cases at the peak, in comparison to the corresponding scenario where natural immunity is not boosted. It should be mentioned that boosting of vaccine-derived immunity (at baseline) plays a more significant role in reducing the average number of new daily cases, in comparison to the corresponding boosting of natural immunity (this can be seen by comparing the corresponding peaks in Figures 4 and 6). In particular, while boosting of vaccine-derived immunity (at baseline) will lead to about 90% reduction in the number of new daily cases at the peak, boosting of natural immunity (at baseline) will lead to about 62% reduction in the number of new daily cases at the peak).

    Further significant reductions in the average number of new daily cases are recorded if the natural immunity wanes at a slower rate (compare the blue and green curves in Figures 6(a) and (c) or (b) and (d), without and with boosting of natural immunity). We illustrated similar dynamics with respect to the cumulative number of new cases without (Figures 6(e) and (f)) and with (Figures 6(g) and (h)) boosting of natural immunity.

    In summary, like in the case of waning of vaccine-derived immunity discussed in Section 5.1, the simulations in this section show that while the waning natural immunity only causes a marginal increase in the average number of new cases at the peak, boosting natural immunity (at baseline) resulted in a significant reduction in the average number of new cases recorded at the peak, in comparison to the scenario where a strategy for boosting of natural immunity is not implemented in the community.

    The COVID-19 pandemic, caused by SARS-CoV-2, has made a significant impact on public health and the economy of almost every nation on earth since its emergence in December of 2019. The United States became the epicenter of the pandemic since late May, 2020 (recording the highest numbers of cumulative cases, hospitalizations and deaths). As of mid July, 2022, the virus had caused over 88.6 million and 1 million deaths in the United States [2,3]. The rapid development, deployment, and administration of several safe and highly effective vaccines contributed significantly in curbing the spread of the virus worldwide. Three of these vaccines (the Pfizer-BioNTech, Moderna and Johnson & Johnson vaccines) have been approved by the FDA for use in the United States. The effectiveness of these vaccines in combating COVID-19 has been negatively affected by the emergence of various variants of SARS-CoV-2 (e.g., the Delta and Omicron variants). In particular, the Omicron (B.1.1.529) variant was declared a variant of concern by the World Health Organization in late November, 2021 [69], due to its exceptionally high transmissibility. Although all the available vaccines were developed for the original SARS-CoV-2 virus strain, they have been able to offer some level of cross-protection against other variants of concern. Furthermore, multiple studies have shown that the efficacy of vaccine-derived immunity wanes over time [28,35,70]. In order to overcome the waning effect of vaccine-derived immunity, booster vaccines were recommended by the CDC in November 2021 [42,71].

    In this study, we developed a mathematical model to assess the population-level impact of the waning and boosting of vaccine-derived and natural immunity against the Omicron BA.1 variant of SARS-CoV-2 in the United States. The model was parameterized by fitting it to the observed cumulative COVID-19 case data for the United States for the period from November, 28, 2021 to February 23, 2022 [3]. We used the remaining segment of the available data (i.e., the segment from February 24, 2022 to March 23, 2022) to cross validate the model. This cross validation, together with simulations involving the new daily COVID-19 cases, showed a good match to the observed data.

    The model was rigorously analyzed to gain qualitative insight into the dynamics and burden of the diseases. The analysis showed that the disease-free equilibrium (DFE) of the model is locally-asymptotically stable whenever the vaccination reproduction number (denoted by Rv<1) is below one. Using the baseline values of the fixed and estimated parameters of the model, we computed the numerical value of Rv during the period of the emergence and circulation of the Omicron variant (starting from late November of 2021). The computed value was Rv=0.81 (suggesting that Omicron was on a downward trajectory towards elimination in the United States). The numerical value of the basic reproduction number of the model (which is computed in the absence of any control measure implemented) was R0=2.051. The computed value of R0 falls within the estimated range of R0 values for Omicron presented in several modeling studies (such as those in [72,73,74,75]). We showed that the disease-free equilibrium of the model is globally-asymptotically for two special cases ((a) when the vaccines offer 100% protection against acquisition of infection and no reinfection and waning of immunity occurs and (b) disease-induced mortality is negligible and reinfection does not occur) when the associated vaccination reproduction number is less than one. The epidemiological implication of this global asymptotic stability result is that the SARS-CoV-2 pandemic can be eliminated if the the associated vaccination reproduction number can be brought to (and maintained at) a value less than one (in other words, having the value of this reproduction threshold less than one is necessary and sufficient for the elimination of the pandemic in the United States).

    Explicit expression for the vaccine-induced herd immunity threshold was derived, and we showed, using current data for COVID-19 cases in the United States, that, for the case where the three vaccines offer 85% average protective efficacy against the Omicron variant, vaccine-derived herd immunity will be achieved in the United States via a combined vaccination-boosting strategy that entails fully-vaccinating 59% of the wholly-susceptible population combined with the boosting of at least 71.5% of the population of the fully-vaccinated individuals whose vaccine-derived immunity has waned to moderate or low level. On the other hand, if the average protective efficacy offered by the three vaccines is reduced to a lower level, such as 55% (as against 85% above), at least 91% of the wholly-susceptible population need to be vaccinated to achieve herd immunity. Furthermore, if the average vaccine efficacy offered by the aforementioned three vaccines drops to 50% (or drops to a much lower value, as presented by Barnard et al. [16], or as stated in several studies on the effectiveness of these vaccines against Omicron [57,72,76,77]), then almost 100% of the wholly-susceptible population need to be vaccinated in order to achieve vaccine-derived herd immunity in the United States. This very high level of vaccination coverage (i.e., vaccinating almost 100% of the susceptible population) is not realistically attainable, especially in large populations such as the United States. Data related to COVID-19 from the United States Centers for Disease Control and Prevention show that, as of May 20, 2022, about 66.5% of the U.S. population was fully-vaccinated and 46.4% of this population received a booster [54]. Thus, our study suggests that, for the scenario that the three vaccines offer the reasonably high average protective efficacy of 85% against acquisition of infection, herd immunity can realistically achieved in the United States by fully-vaccinating a moderate proportion (about 59%) of the wholly-susceptible and boosting about 71.5% of this cohort in whom the vaccine-derived immunity has waned to moderate or low level.

    We conducted extensive numerical simulations to assess the impact of waning and boosting of vaccine-derived and natural immunity for each three arbitrarily selected waning scenarios (slow, baseline, and fast). Our study showed, based on these simulations, that in the absence of boosting of vaccine-derived and natural immunity, waning of vaccine-derived and natural immunity only causes a marginal increase in the average number of daily cases (at the peak) and the number of cumulative COVID-19 cases, in comparison to the baseline scenario. In other words, we showed that waning of either vaccine-derived or natural immunity (or both) has only marginal impact, for each of the three waning scenarios we considered, on the dynamics of the SARS-CoV-2 pandemic (as measured in terms of increases in the average number of daily new cases recorded at the peak, in comparison to the case where baseline values of all the parameters of the model are used).

    We also showed that if fully-vaccinated individuals with moderate or low level of vaccine-derived immunity are boosted (at baseline level), the effect of waning of immunity is a lot less pronounced, in comparison to the baseline scenario (in other words, dramatic reductions in the increase in the average number of daily new cases at the peak recorded (under the three waning scenarios) are achieved if both immunity types are boosted at baseline level, in comparison to the corresponding scenarios where the immunity wanes but no boosting is implemented. We further showed that boosting of vaccine-derived immunity is more beneficial (in reducing average number of new cases) than boosting of natural immunity. Specifically, for the fast waning scenario, boosting of vaccine-derived immunity (at baseline level) resulted in an approximate 90% reduction in the average number of new daily cases at the peak, while boosting of natural immunity resulted in about 62% reduction in the number of new daily cases at the peak (in comparison to the corresponding scenarios without boosting).

    Furthermore, this study shows that boosting of vaccine-derived immunity (implemented near the baseline level) increased the prospects of altering the trajectory of the COVID-19 pandemic from persistence to possible elimination (even for the fast waning scenario of the vaccine derived-immunity) of the pandemic in the United States. Thus, the implementation of vaccination-boosting strategy greatly enhances the prospects of eliminating the COVID-19 pandemic in the United States.

    In addition to the standard assumptions on which the model is built, some of the limitations of this study include the fact that we did not explicitly account for the impact of other control interventions (notably, the use of face masks, voluntary testing and detection of SARS-CoV-2 cases, isolation of confirmed cases, etc.), which also play important roles in the battle against the COVID-19 pandemic. Furthermore, this study assumes that the population is well-mixed and does not explicitly account for a number of heterogeneities, including age and risk structure, which may be relevant to gain insight into the dynamics of the disease. Furthermore, the current study did not account for the effects of other SARS-CoV-2 variants, including the BA.2 Omicron variant (which is more contiguous than the original BA.1 Omicron variant) [19,26]. We fitted our deterministic model to cumulative case data instead of raw (new daily case) data, which might lead to narrower confidence intervals and/or provide a misleading measure of uncertainty. Hence, the results of the study should be interpreted with some caution. Nonetheless, our study shows, overall, that the prospect for the effective control and mitigation (and, consequently, elimination) of the COVID-19 pandemic in the United States is very promising using a combined vaccination-boosting strategy, provide the vaccinate and boosting coverages are moderately high enough.

    ABG acknowledges the support, in part, of the Simons Foundation (Award #585022) and the National Science Foundation (Grant Number: DMS-2052363). CNN acknowledges the support of the Simons Foundation (Award #627346) and the National Science Foundation (Grant Number: DMS #2151870). SS acknowledges the support of the Fulbright Foreign Student Program. The authors are very grateful to the three anonymous reviewers for their very constructive comments.

    The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.

    Figure A1.  Complete and connected flow diagram of the model (2.1).

    Proof. Consider the special case of the model (2.1) with δp=δs=δa=δh=0 and εni=εnvi=1 (i=1,2,3). Further, let ˆRv<1. The equations for the infected compartments of this special case of the model can be re-written in terms of the next generation matrices (ˆF and ˆV) as follows:

    ddt[E(t)Ip(t)Is(t)Ia(t)Ih(t)]=(ˆFˆV)[E(t)Ip(t)Is(t)Ia(t)Ih(t)]ˆM[E(t)Ip(t)Is(t)Ia(t)Ih(t)], (B.1)

    where (with S and A as defined in Section 3),

    (ˆFˆV)=[(σE+μ)βηp(S+A)βηs(S+A)βηa(S+A)βηh(S+A)σE(σp+γp+μ)0000rσp(ϕs+γs+μ)000σp(1r)0(γa+μ)000ϕs0(γh+μ)],

    and,

    ˆM=β[(SS)+(1εv1)(V1V1)+(1εv2)(V2V2)+(1εv3)(V3V3)][0ηpηsηaηh00000000000000000000]. (B.2)

    Since SS, V1V1, V2V2 and V3V3 for all t>0 in Ω, it follows that the matrix ˆM, defined in (B.2), is non-negative. Hence, Eq (B.1) can be re-written in terms of the following inequality:

    ddt[E(t)Ip(t)Is(t)Ia(t)Ih(t)](ˆFˆV)[E(t)Ip(t)Is(t)Ia(t)Ih(t)]. (B.3)

    The proof is concluded the same way as in the proof of Theorem 3.2. Thus, the DFE (E0) of the special case of the model (2.1) (with δp=δs=δa=δh=0 and εni=εnvi=1 (i=1,2,3)) is globally-asymptotically stable in Ω whenever ˆRv<1.



    [1] A. Almeida, D. Drihem, Maximal, potential and singular type operators on Herz spaces with variable exponents, J. Math. Anal. Appl., 394 (2012), 781–795. https://doi.org/10.1016/j.jmaa.2012.04.043 doi: 10.1016/j.jmaa.2012.04.043
    [2] S. Albeverio, A. Y. Khrennikov, V. M. Shelkovich, Harmonic analysis in the p-adic Lizorkin spaces: fractional operators, pseudo-differential equations, p-adic wavelets, Tauberian theorems, J. Fourier Anal. Appl., 12 (2006), 393–425. https://doi.org/10.1007/s00041-006-6014-0 doi: 10.1007/s00041-006-6014-0
    [3] M. Z. Baber, N. Ahmed, C. J. Xu, M. S. Iqbal, T. A. Sulaiman, A computational scheme and its comparison with optical soliton solutions for the stochastic Chen-Lee-Liu equation with sensitivity analysis, Mod. Phys. Lett. B, 2024 (2024), 2450376. https://doi.org/10.1142/S0217984924503767 doi: 10.1142/S0217984924503767
    [4] C. Capone, D. Cruz-Uribe, A. Fiorenza, The fractional maximal operator and fractional integrals on variable Lp spaces, Rev. Mat. Iberoamericana, 23 (2007), 743–770. https://doi.org/10.4171/RMI/511 doi: 10.4171/RMI/511
    [5] D. V. Cruz-Uribe, A. Fiorenza, Variable Lebesgue spaces: foundations and harmonic analysis, Basel: Springer, 2013. https://doi.org/10.1007/978-3-0348-0548-3
    [6] L. F. Chacón-Cortés, H. Rafeiro, Variable exponent Lebesgue spaces and Hardy–Littlewood maximal function on p-adic numbers, P-Adic Num. Ultrametr. Anal. Appl., 12 (2020), 90–111. https://doi.org/10.1134/S2070046620020028 doi: 10.1134/S2070046620020028
    [7] N. M. Chuong, H. D. Hung, Maximal functions and weighted norm inequalities on local fields, Appl. Comput. Harmon. A., 29 (2010), 272–286. https://doi.org/10.1016/j.acha.2009.11.002 doi: 10.1016/j.acha.2009.11.002
    [8] R. Coifman, C. Fefferman, Weighted norm inequalities for maximal functions and singular integrals, Stud. Math., 51 (1974), 241–250.
    [9] B. Dragovich, A. Y. Khrennikov, S. V. Kozyrev, I. V. Volovich, On p-adic mathematical physics, P-Adic Num. Ultrametr. Anal. Appl., 1 (2009), 1–17. https://doi.org/10.1134/S2070046609010014 doi: 10.1134/S2070046609010014
    [10] K. H. Dung, D. V. Duong, The p-adic Hausdorff operator and some applications to Hardy–Hilbert type inequalities, Russ. J. Math. Phys., 28 (2021), 303–316. https://doi.org/10.1134/S1061920821030043 doi: 10.1134/S1061920821030043
    [11] K. H. Dung, D. V. Duong, Two-weight estimates for Hardy–Littlewood maximal functions and Hausdorff operators on p-adic Herz spaces, Izv. Math., 87 (2023), 920–940. https://doi.org/10.4213/im9404e doi: 10.4213/im9404e
    [12] K. H. Dung, D. L. C. Minh, T. T. Nang, Boundedness of Hardy–Cesàro operators on variable exponent Morrey–Herz spaces, Filomat, 37 (2023), 1001–1016. https://doi.org/10.2298/FIL2304001D doi: 10.2298/FIL2304001D
    [13] K. H. Dung, P. T. K. Thuy, Commutators of Hardy–Littlewood operators on p-adic function spaces with variable exponents, Open Math., 21 (2023), 20220579. https://doi.org/10.1515/math-2022-0579 doi: 10.1515/math-2022-0579
    [14] L. Diening, M. Ružička, Calderón-Zygmund operators on generalized Lebesgue spaces Lp(x) and problems related to fluid dynamics, J. Reine Angew. Math., 563 (2003), 197–220. https://doi.org/10.1515/crll.2003.081 doi: 10.1515/crll.2003.081
    [15] C. Fefferman, E. M. Stein, Some maximal inequalities, Am. J. Math., 93 (1971), 107–115. https://doi.org/10.2307/2373450 doi: 10.2307/2373450
    [16] J. Garcìa-Cuerva, E. Harboure, C. Segovia, J. L. Torrea, Weighted norm inequalities for commutators of strongly singular integrals, Indiana U. Math. J., 40 (1991), 1397–1420.
    [17] L. Grafakos, Modern Fourier analysis, New York: Springer, 2008. https://doi.org/10.1007/978-1-4939-1230-8
    [18] Q. J. He, X. Li, Necessary and sufficient conditions for boundedness of commutators of maximal function on the p-adic vector spaces, AIMS Mathematics, 8 (2023), 14064–14085. https://doi.org/10.3934/math.2023719 doi: 10.3934/math.2023719
    [19] M. Izuki, Fractional integrals on Herz–Morrey spaces with variable exponent, Hiroshima Math. J., 40 (2010), 343–355. https://doi.org/10.32917/hmj/1291818849 doi: 10.32917/hmj/1291818849
    [20] M. Izuki, T. Noi, Two weighted Herz spaces with variable exponents, Bull. Malays. Math. Sci. Soc., 43 (2020), 169–200. https://doi.org/10.1007/s40840-018-0671-4 doi: 10.1007/s40840-018-0671-4
    [21] A. Khrennikov, p-Adic valued distributions in mathematical physics, Dordrecht: Springer, 1994. https://doi.org/10.1007/978-94-015-8356-5
    [22] A. N. Kochubei, Radial solutions of non-Archimedean pseudodifferential equations, Pacific Jounal of Mathematics, 269 (2014), 355–369. https://doi.org/10.2140/pjm.2014.269.355 doi: 10.2140/pjm.2014.269.355
    [23] S. V. Kozyrev, Methods and applications of ultrametric and p-adic analysis: From wavelet theory to biophysics, Proc. Steklov Inst. Math., 274 (2011), 1–84. https://doi.org/10.1134/S0081543811070017 doi: 10.1134/S0081543811070017
    [24] Y. C. Kim, Lq-Estimates of maximal operators on p-adic vector space, Commun. Korean Math. S., 24 (2009), 367–379. https://doi.org/10.4134/CKMS.2009.24.3.367 doi: 10.4134/CKMS.2009.24.3.367
    [25] D. F. Li, G. E. Hu, X. L. Shi, Weighted norm inequalities for the maximal commutators of singular integral operators, J. Math. Anal. Appl., 319 (2006), 509–521. https://doi.org/10.1016/j.jmaa.2005.06.054 doi: 10.1016/j.jmaa.2005.06.054
    [26] P. L. Li, R. Gao, C. J. Xu, J. W. Shen, S. Ahmad, Y. Li, Exploring the impact of delay on Hopf bifurcation of a type of BAM neural network models concerning three nonidentical delays, Neural Process. Lett., 55 (2023), 11595–11635. https://doi.org/10.1007/s11063-023-11392-0 doi: 10.1007/s11063-023-11392-0
    [27] S. Z. Lu, D. C. Yang, The decomposition of weighted Herz space on Rn and its applications, Sci. China. Ser. A, 38 (1995), 147–158.
    [28] S. Z. Lu, D. C. Yang, The decomposition of Herz spaces on local fields and its applications, J. Math. Anal. Appl., 196 (1995), 296–313. https://doi.org/10.1006/jmaa.1995.1411 doi: 10.1006/jmaa.1995.1411
    [29] Y. Mizuta, T. Ohno, T. Shimomura, Boundedness of maximal operators and Sobolev's theorem for non-homogeneous central Morrey spaces of variable exponent, Hokkaido Math. J., 44 (2015), 185–201. https://doi.org/10.14492/hokmj/1470053290 doi: 10.14492/hokmj/1470053290
    [30] K. S. Rim, J. Lee, Estimates of weighted Hardy–Littlewood averages on the p-adic vector space, J. Math. Anal. Appl., 324 (2006), 1470–1477. https://doi.org/10.1016/j.jmaa.2006.01.038 doi: 10.1016/j.jmaa.2006.01.038
    [31] C. Segovia, J. L. Torrea, Higher order commutators for vector-valued Calderón–Zygmund operators, T. Am. Math. Soc., 336 (1993), 537–556. https://doi.org/10.2307/2154362 doi: 10.2307/2154362
    [32] E. M. Stein, Harmonic analysis, real-variable methods, orthogonality, and oscillatory integrals, Princeton: Princeton University Press, 1993.
    [33] M. Sultan, B. Sultan, A. Aloqaily, N. Mlaiki, Boundedness of some operators on grand Herz spaces with variable exponent, AIMS Mathematics, 8 (2023), 12964–12985. https://doi.org/10.3934/math.2023653 doi: 10.3934/math.2023653
    [34] J. Tan, Boundedness of multilinear fractional type operators on Hardy spaces with variable exponents, Anal. Math. Phys., 10 (2020), 70. https://doi.org/10.1007/s13324-020-00415-x doi: 10.1007/s13324-020-00415-x
    [35] V. S. Vladimirov, I. V. Volovich, p-Adic quantum mechanics, Commun. Math. Phys., 123 (1989), 659–676. https://doi.org/10.1007/BF01218590 doi: 10.1007/BF01218590
    [36] V. S. Vladimirov, I. V. Volovich, E. I. Zelenov, p-Adic analysis and mathematical physis, Singapore: World Scientific, 1994. https://doi.org/10.1142/1581
    [37] S. S. Volosivets, Multidimensional Hausdorff operator on p-adic field, P-Adic Num. Ultrametr. Anal. Appl., 2 (2010), 252–259. https://doi.org/10.1134/S2070046610030076 doi: 10.1134/S2070046610030076
    [38] S. S. Volosivets, Maximal function and Riesz potential on p-adic linear spaces, P-Adic Num. Ultrametr. Anal. Appl., 5 (2013), 226–234. https://doi.org/10.1134/S2070046613030059 doi: 10.1134/S2070046613030059
    [39] H. Wang, The decomposition for the Herz spaces, Pacific Jounal of Mathematics, 25 (2015), 15–28.
    [40] C. J. Xu, W. Ou, Q. Y. Cui, Y. C. Pang, M. X. Liao, J. W. Shen, et al., Theoretical exploration and controller design of bifurcation in a plankton population dynamical system accompanying delay, Discrete Cont. Dyn.-S, 2024 (2024), 36. https://doi.org/10.3934/dcdss.2024036 doi: 10.3934/dcdss.2024036
    [41] C. J. Xu, Y. Y. Zhao, J. T. Lin, Y. C. Pang, Z. X. Liu, J. W. Shen, et al., Bifurcation investigation and control scheme of fractional neural networks owning multiple delays, Comp. Appl. Math., 43 (2024), 186. https://doi.org/10.1007/s40314-024-02718-2 doi: 10.1007/s40314-024-02718-2
    [42] C. J. Xu, J. T. Lin, Y. Y. Zhao, Q. Y. Cui, W. Ou, Y. C. Pang, et al., New results on bifurcation for fractional-order octonion-valued neural networks involving delays, Network-Comp. Neural, 2024 (2024), 1–53. https://doi.org/10.1080/0954898X.2024.2332662 doi: 10.1080/0954898X.2024.2332662
    [43] T. L. Yee, K. L. Cheung, K. P. Ho, C. K. Suen, Local sharp maximal functions, geometrical maximal functions and rough maximal functions on local Morrey spaces with variable exponents, Math. Inequal. Appl., 23 (2020), 1509–1528. https://doi.org/10.7153/mia-2020-23-108 doi: 10.7153/mia-2020-23-108
  • This article has been cited by:

    1. Hemaho B. Taboe, Michael Asare-Baah, Enahoro A. Iboi, Calistus N. Ngonghala, Critical assessment of the impact of vaccine-type and immunity on the burden of COVID-19, 2023, 00255564, 108981, 10.1016/j.mbs.2023.108981
    2. Carly Meyer, Louis Goffe, Vivi Antonopoulou, Fiona Graham, Mei Yee Tang, Jan Lecouturier, Aikaterini Grimani, Paul Chadwick, Falko F Sniehotta, Using the precaution adoption process model to understand decision-making about the COVID-19 booster vaccine in England, 2023, 0264410X, 10.1016/j.vaccine.2023.02.047
    3. Adel Alatawi, Abba B. Gumel, Mathematical assessment of control strategies against the spread of MERS-CoV in humans and camels in Saudi Arabia, 2024, 21, 1551-0018, 6425, 10.3934/mbe.2024281
    4. Salman Safdar, Abba B. Gumel, 2023, Chapter 10, 978-3-031-40804-5, 243, 10.1007/978-3-031-40805-2_10
    5. Ugo Avila-Ponce de León, Angel G. C. Pérez, Eric Avila-Vales, Modeling the SARS-CoV-2 sublineages XBB and BQ.1 in Mexico, considering multiple vaccinations, booster dose, waning immunity and cross-immunity, 2024, 32, 2688-1594, 1082, 10.3934/era.2024053
    6. Chidozie Williams Chukwu, Zviiteyi Chazuka, Salman Safdar, Iffatricia Haura Febriana, Dipo Aldila, Assessing Syphilis transmission among MSM population incorporating low and high-risk infection: a modeling study, 2024, 43, 2238-3603, 10.1007/s40314-024-02669-8
    7. Yi Yin, Shuhan Tang, Qiong Li, Sijia Zhou, Yuhang Ma, Weiming Wang, Daihai He, Zhihang Peng, Estimate the number of lives saved by a SARS-CoV-2 vaccination campaign in six states in the United States with a simple model, 2024, 12, 27727076, 100390, 10.1016/j.ijregi.2024.100390
    8. Ugo Avila-Ponce de León, Angel G. C. Pérez, Eric Avila-Vales, Modeling the SARS-CoV-2 Omicron variant dynamics in the United States with booster dose vaccination and waning immunity, 2023, 20, 1551-0018, 10909, 10.3934/mbe.2023484
    9. Queen Tollett, Salman Safdar, Abba B. Gumel, Dynamics of a two-group model for assessing the impacts of pre-exposure prophylaxis, testing and risk behaviour change on the spread and control of HIV/AIDS in an MSM population, 2024, 9, 24680427, 103, 10.1016/j.idm.2023.11.004
    10. Binod Pant, Salman Safdar, Calistus N. Ngonghala, Abba B. Gumel, Mathematical Assessment of Wastewater-Based Epidemiology to Predict SARS-CoV-2 Cases and Hospitalizations in Miami-Dade County, 2025, 73, 0001-5342, 10.1007/s10441-025-09492-6
  • Reader Comments
  • © 2024 the Author(s), licensee AIMS Press. This is an open access article distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/4.0)
通讯作者: 陈斌, bchen63@163.com
  • 1. 

    沈阳化工大学材料科学与工程学院 沈阳 110142

  1. 本站搜索
  2. 百度学术搜索
  3. 万方数据库搜索
  4. CNKI搜索

Metrics

Article views(931) PDF downloads(48) Cited by(1)

Other Articles By Authors

/

DownLoad:  Full-Size Img  PowerPoint
Return
Return

Catalog