Scribd Logo
Upload

Welcome to Scribd!

  • A Model Based Study On The Dynamics of COVID-19 - Prediction and Control - Elsevier Enhanced Reader

    Uploaded by

    genius
    8 views12 pages

    Original Title

    A model based study on the dynamics of COVID-19_ Prediction and control _ Elsevier Enhanced Reader

    Copyright

    © © All Rights Reserved

    Available Formats

    PDF or read online from Scribd

    Did you find this document useful?

    Is this content inappropriate?

    Report this Document

    Copyright:

    © All Rights Reserved
    Download as PDF or read online from Scribd
    Flag for inappropriate content
    8 views12 pages

    A Model Based Study On The Dynamics of COVID-19 - Prediction and Control - Elsevier Enhanced Reader

    Uploaded by

    genius

    Copyright:

    © All Rights Reserved
    Download as PDF or read online from Scribd
    Flag for inappropriate content
    of 12
    (aos, Stone and fracas 196 (2020 T0N889 Contents sts available at ScienceDirect Chaos, Solitons and Fractals Nonlinear Science, and Nonequilibrium and Complex Phenomena journal homepage: www elseviercomlocatelchaos A model based study on the dynamics of COVID-19: Prediction and ® control oan Manotosh Mandal’, Soovoojeet Jana", Swapan Kumar Nandi‘, Anupam Khatua®, Sayani Adak", TK. Kar‘ *Orparime’ of Matemati,Tonsite Makoto, Tm ER Wes ea na Deparment of Maremat’: Rasy Cope ft Hoa 74, es Berle ‘Nyasa PL Sishaetan,Paschim Medmpr 721253, West Bega as "Deparment of Mathomais, Indion nse of Engnering Se and Tete, Ship, Howrah 7103, io ARTICLE INFO ABSTRACT ‘ace an Reeve 28 api 2020, Revised 6 May 2020, Deceeed 10 Bay 2020 ‘aa eine 13 May 2020 {as there fe no vaccination and proper medicine for treatment, the resent pandemic caused by COVID- 19 has drawn attention tothe strategies of quarantine and other governmental measures, ik lockdown, media coverage on socal isolation, and improvement of public hygiene, ete to contol the disease. The mathematical model can help when these intervention measures ae the best strateges for disease con twol as well as hove they might affect the disease dynamics. Motivated by this, in this article. we have formulated a mathematical model introducing a quarantine cass and governmental intervention mea- Theoretical epidemitoay sures f0 mitigate disease transmission. We study a thorough dynamical behavior of the model in texms Basi eration number ofthe basic repredction number, Further, we perform the sensitivity analysis ofthe essential reproduc ee tion number and found tha reducing the contact of exposed and susceptible humans isthe most critical Bangg a gar contol factor in achieving disease contro. To lessen the infected individuals 25 well as t0 minimize the cost eon of implementing government contol measures, we formulate an optinal contol problem, nx optimal contol is determined. Finally, we forecast short-term tend of COVID-I8 for the theee highly affected states, Maharashtra, Delhi and Tamil Nadu, in India, and it suggests tha the fst two states need farther poor ‘monitoring of control measures to reduce the contact of exposed and susceptible humans (© 2020 Elsevier Lk, All rights reserve. 1. Introduction ‘The recent pandemic, which is commonly known as COVID-19, is an infectious disease caused by the virus, severe acute respira- tory syndrome coronavirus 2 (SARS-CoV-2). In late December 2018, the disease COVID-19 was first identified in Wuhan, the capital Hubei province. China, and causing the first pandemic of this cen- tury. Since the stat af the year 2020, the infectious disease COVID- 19 has started to spread globally and resulting in almost three mil lion positive cases til today (cll 27th April, 2020 there with about ‘one million closed cases having a 20% death rate). As we under- stand, the disease COVID-19 has become a pandemic due to sev- ‘eral reasons. Some of them are (i) unavailability of both suitable vaccine and exact medicine, (ii) high disease transmission rate, Conesponding author Erma addressee manetosnnsgmailcom (M, Manda, soovsjeregmslcom (S. Jana), nandismapensoogmaiicom (SK. Nandi akhatuads@gmaicom (A ata), ayniadak940ymacom (Ad) ar 117Bgmcom (TKK ps aatorg)11016}.4305.2020105889 (960-0779) 2090 Eee UA sighs veserved, and (i) the precise nature of the virus SARS-CoV-2 being stil un= known, As many as 210 countries and territories around the globe have been suffering fram COVID-19 today [see [16]. ‘The primary step taken by most of the governmental agencies to control COVID-19 is the implementation of lockdown to main- tain social distance. This procedure is an excellent measure to con- trol the spreading of the disease, Still, from an economic point ‘of view, the complete lockdown may be the cause of a signifi. ‘ant financial crisis for the near future, In particular, lockdown in high dense countries may reduce the disease transmission rate, al- though complete control may not be achievable. Hence to alive the ‘economic status of a country, a full fockdown for an indefinite pe- riod is not desirable at all in any circumstances. Therefore there should be a suitable balance between the two different character- istics of governmental policies complete lockdown and healthy free conditions. ‘Therefore, in the present scenario, a qualitative analysis of COVID-19 is more significant compare to quantitative analysis Hence 2 suitable mathematical model would not only able to 2 -M. Mondat, Jana and SK Noa cto Chas, Sotons and Fractal 1952020) 10889 represent the whole disease system but also the study of the ‘model would undoubtedly derive the precise nature of the dis- cease, It may forecast the behavioral aspect of the disease shorty, Although the primitive mathematical models on theoretical epi- demiology (see Bernoulli |, Hamer [13], Ross [40]. Kermack and MeKendric [24], ete) look quite simple, stil, from a new perspec- tive, those works are the major works on mathematical epidemiol- ogy. With the advancement of computational tools and software, we can develop a complex mathematical model and analyze it thoroughly in a scientific manner. In the history of literature. we see that many model-based studies have successfully achieved the slobal dynamics of the corresponding infectious disease (see the references Keeling and Rohani [23], Wang and Zhao |, Buonomo et al. [3], Zhou et al. [43], Jana et al. [18.20], Li etal. [29), Zegarra and Hernandez (42| etc) ‘According to the information received, it may take around one week to two weeks for the exposure of symptoms of COVID-19 of am infected person, although during this period, that person able to Infect other susceptible persons. However, there may be some i fected persons whose infection is so mild that the person would recover due to innate immunity even before the hospitalization. Thus in this article, by the term ‘infected’ person, we will mean those persons who are hospitalized. Further, we assume that the ‘medical personals assisting COVID-19 positive hospitalized individ- uals have taken necessary protective items. Thus to keep simplic- ity, we believe that only exposed persons and asymptomatic it fected persons can spread the disease. Stil, the infected persons Who are hospitalized in due course are not spreading the virus. Al though the effect of COVID-19 started not more than six months ‘ago, some significant works have already been done of ate ut {ergoing to expose COVID-19 with the help of suitable mathemat- ical models. Kucharski et al. [25] have proposed a mathematical ‘model-based analysis of COVID-19, where the authors have consid- ered all the positive cases of Wuhan, China, till Sth March 2020. In another recent work, Ndairou et al. [35] have presented an au- onomous system of mathematical model to study the spread of COVID-19 in the Wuhan city, China. Using mathematical modeling Prem et al. [38] have analyzed the controlling status of COVID- of Wuhan city. Hellewell et al (15] have described the effective procedure of COVID-19 disease using isolation. In another work, “Mizumoto and Chowell [22) have studied the transmission dynam ies of COVID-19 in the international conveyance, Diamond Princess Cruises Ship. The basic reproduction number is an important pa- rameter to analyze the nature of an infectious disease. In their Work, Liu eta. [30] have investigated the possible numerical value of the basie reproduction number of COVID-19. In a recent aticle, Fanelli and Piazza (7) have analyzed and predicted the nature of COVID-19 in three most affected countries till March 2020 with the help of mathematical modelling. Ribeizo etal [39] have used some stochastic based regression models to forecast the phenomena in as many as ten most affected states of Brazil In other recent work Chakraborty and Ghosh [4 have considered a hybrid ARIMA-WBF ‘model to forecast various COVID-19 affected countries throughout the globe. Since the nature and destruction of COVID-19 depends on vari- us parameters (namely personal immunity, history of visiting into 2 COVID-19 pandemic country, maintaining the required hygiene etc) of the affected system, using a single model we cannot de- scribe the whole disease system throughout the globe. Addition of detail and complexity can make models more accurate but ths also complicates their mathematics. However, even this kind of simple ‘model is very helpful when formal vaccination or proper treatment control isnot available, Motivated by this. in the present paper, we propose a mathematical model introducing a quarantine class and governmental intervention measures like lockdown, media cover age on social distancing, and improvement of public hygiene, ete {e mitigate disease transmission. The main purpose of this work is to explore the role of quarantine and the governmental interven- tion strategies on COVID-19 control and elimination. We study a thorough dynamical behavior of the model in terms of the basic ‘reproduction number. Further, we perform the sensitivity analysis of the essential reproduction number and found that reducing the contact of exposed and susceptible humans is the most critical fac {or in achieving disease control. To lessen the infected individuals as well as to minimize the cost of implementing government con- {rol measures, we formulate an optimal control problem, and op- timal control is determined, Finally, we predict a short-term trend ‘of COVID-19 for the three most affected states, Mahatashtra, Delhi, and Tamil Nadu in India, and it suggests that the fist two states need further monitoring of control measures to reduce the contact of exposed and susceptible humans, The rest portion of this paper has been organized in the follow- ing way: In Section 2, we describe the theoretical analysis of the article, which includes the basic reproduction number, existence, and asymptotic stability criteria of the two equilibria, namely disease-free and endemic equilibrium point. In Section 3, we for ‘mulate an optimal control problem and solve the problem analyt- ically. In Section 4, we provide some numerical examples to ilus- trate the theoretical counterparts. Next, in Section 5, we consider seal-world examples on COVID-19 of three most COVID-19 affected states in India and derive some short term forecasting based on the proposed model. In the final section, we provide a thorough discussion of our article. 2, Model formulation In this section, we formulate a mathematical model on COVID- 19 based on some realistic assumptions. At any time instant f, the human populations are subdivided into five time-dependent classes, namely Susceptible S(t), Exposed F(0), Hospitalized infected Ie), Quarantine Q(t) and Recovered or Removed R(t). Based on those five state variables, we aim to form an autonomous system using first-order differential equations. Let A be the constant re- cruitment rate to the susceptible population, and 8 be the disease twansmission rate. However the disease transmission from vulner- able to infected persons (here the class E) depends on several pa- rameters, namely, precautions (use of face mask, social distancing, not rubbing face and nose using hand, etc.) and hygienic environ- ‘ment (use of soap and sanitizer, hand washing, cleaning, et. taken by both susceptible as well as infected persons, Since here, we have assumed that the virus of COVID-I9 is spreading when a vul- erable person comes into contact with an expased person; there- fore we think that p,(0 0 4 Analysis of the system for fxed control ln this section we assume the fixed value for control parameter M, Here we mainly study the uniformly boundedness of the so- lutions and subsequently the basic reproduction number, different ‘equilibria and their stability criteria, sensitivity ete. 3.1. Boundedness ofthe system Here we examine the boundedness property of the system (1). ‘Theorem 3:1. The solutions of the system (1) are uniformly bounded, Proof. We assume that X=S + FQ 4148 $ a Therefore ax qt il=A—dx ax ie, Rd A Integrating the above inequality and by applying the theorem of differential equation due to Birkhoff and Rota [2], we get [1] 4x0 Tow Now for ¢ > c, A OS NS Tpit Hence all the solutions of (1) that are initiating in {R§) are con- fined in the region Ate Tp for any ¢ > 0 and for t+ co, Hence the theorem, 0) (Ke RE0=XE QR) < 32, Basie reproduction mumer In any epidemic model, the basic reproduction number is the ‘most important epidemiological parameter for determining the na~ ture of a disease. Generally i is denoted by Ry and is defined as the “number of secondary infected individuals caused by single in- fected individuals in the whole time interval” (see, van den Driess- ‘che and Watmough [6]. Therefore the dimensionless quantity Ro refers as the expectation of the spreading disease. In literature, several techniques ate available to evaluate Ry for ‘an epidemic spread. In our present research article we use the next 4 -M. Mondat, Jana and SK Noa cto Chas, Sotons and Fractal 1952020) 10889 generation matrix approach (5,922), Now the classes which are di rectly involved for spread of disease is only &, Q I. Therefore from system (1) we have & — pa py past - nsf -et -0F de dQ 42 _ 5.6 m9 coo @ a Hak sansa soo. The above system can be writen as $= O09) - WO). p Athan) wee y=(a). on (" "8 w= i ° (taro+ae (by +e+d)Q— Doe (y 4d +S) 0k—eQ is clear from the system wat Eo AG. atts It is clear from the system (1) that Eo(zAqy.0.0.0. i) is a disease free equilibrium. Now the Jacobian matrix of «> and W at the disease free equilibrium are respectively given by, (BC — pr pn)" 00} (O16) = ° 0 0) and V=J(¥ Eo) ° 0 9 neatord 0 0 =p, btcrd 0 ma -¢ nsdsd The basic reproduction number (Ro) isthe spectral radius of the of the matrix (FV-') and for the present model itis given by AB ~ pil = pa) 80 ip tava +a) @) 33. Equilibria The system has two possible equilibria, One is. disease free equilibrium point. where infection vanishes from the sys- tem. It is given by Eo(S", 0, 0, 0, 8), where $= zy FO = zfiMg,. The other equilibrium point is ES", , 1h) whese infection i abeays present in the system i called endemic equilibria, where S* = 2M Fo = (by be peptone ea ereeatirseaT Dene i ea pain p= lets cB peo egh ig ad) Note I observed from the expression of the above tw equ librium point is that te disease fee equlinum F aways ea sible bt the endemic equilibrium Ey i Tease I Mp 1 34, Stability analysis In the present section we investigate the local asymptotic sta- bility criteria of the different equilibria, Theorem 32. The disease free equilibrium Ey is locally asymptotically stable if Ry < 1 and it is unstable if Ra > 1 Proof. The Jacobian matrix at the disease free equilibrium of the system (1) is given by asp) At pyyt pas? » oo 0” papier? -Grtare+e 0 oO i] a tte) 00 it : ° ow [Now the characteristic equation of the system (1) at its disease free equilibrium is given by C+D +d+ PMV +b) He+ d+ +d-45) C4 (bp bar to +d) Ro) 4) Clearly all the eigen value of the Jacobian matrix are negative Jf and only if Ry = 1. Hence the system is locally asymptotically stable if Ry = 1 and itis unstable if Rp > 1. Hence the theorem. Note Here we see that the disease free equilibrium Ep losses its stability when the Ry increases to its value greater than 1. So, we ‘may conclude that at Rp the system (1) passes through a bifurca- ‘ion around its disease free equilibrium which are discussed in the next theorem, ‘Theorem 3.3. The system (1) passes trough a eranseritica bifurca- tion around its disease free equilibrium whet! Ro = 1 Proof. From the above analysis, it has been observed that when Rp = 1 between the two equilibria, only the disease free equi- librium exists and locally asymptotically stable where as Ro > 1 fs the threshold condition for both existence and asymptotic sta- bility criteria of the endemic equilibrium point although the dis ease free equilibrium reduces to unstable nature at the threshold Rp > |. Hence we may conclude that there is change of feasibility as well as stability occurs at Rg = 1. Following the articles Guken- hheimer and Holmes 12), Kar and Jana 21) Jana etal. (19) ee, itis concluded that the system undergoes a transcrtical bifureation at y= 1 and in Fig. 4, we represent graphically the phenomenon of ‘wanseviticl bifurcation, Now we study the local asymptotic stability of the endemic equilibrium £} Theorem 3.4, The Endemic equilibrium F; is locally asymptotically stable for Ro = Proof. The jacobian matrix for the system (1) is given by fy Bp pass oo pet pte on 0 ° ° by citer 00 ° a © ~rdes) 0 ou ° o nd where, an =—8C~ py) — p2)E*— (d+ PM), a = BO py) = p2)S* = (bao +d), The characteristic equation of the system (1) around its endemic equilibrium E, is 4A +m 444007 462 4GR46)=0 6) where ¢, =2d-+b) +64 PM + BU px) pa) C= (by + 6+ 0)(d-+ pM + BU — py) — pa)E") + (by tet © + )BA — p)(1 ~ p29). C= [bp Feta +d), +64 d) — DyDaI BA — py) ~ pF It is clear from the equation (5) that first two roots are nega~ tive real numbers and remaining roots are the roots of the cubic polynomial. It is also observe that here Cy, Cz, C3 and GC) ~C3 all are postive for any parametric value. Hence following the Routh- Hurwitz criterion we may conclude that the system (1) is locally asymptotically stable around its endemic equilibrium Ey. 0 4, Optimal control problem Here we focus on the time varying control M{t), which repre- sents the awareness due to media coverage. Now it is very impor- {ant to find out a strategy which minimizes the number of infected persons as well as the associated cost. In this regard, the optimal control theory is a very powerful tool to figure out such policy. ‘Therefore we consider the optimal control problem to minimize 1M Mando Jona an SK None at/Chas Solitons and Frac 136 (202) 105880 5 the objective functional. Following [27.31], we construct the objec- tive funetional as follows: if "Teil + coho) 6) ‘subject to the proposed model (1), The parameters cy and ¢2 cor responds as the weight constraints for the infected population and the control respectively. Here the objective functional is linear in the control with bounded states. Therefore it can be be showed by using standard results that an optimal control and corresponding ‘optimal states exist (S|. Now we need to find out the value of the ‘optimal control BP) such that JO") = ming) Where @=(M(Q):0 0, Now to investigate the singular case, we assume that ‘on some non-trivial interval. n this case we caleulate 2 (3) After some simplifications of the time derivative of sf. we ob- tain a (aH) _a@ 5(au) Feat Os baypS) 2 pS — Ay p+ FspS + eps. (12) 22) «assay By + (A—KSE + nQ — d5)(A5 — 21) + pdSis, (13) ‘occur in the above expression, so next we calculate the second ‘desivative with respect to time. + 8G) = [W(E(i2 9a) + 02 — ae) a (REQ — a) — dap + (A= R(SE +5) +10 — 08) sap F(A KSE+ byQ — d5)(%s —7a)p-+ pdSHs + pdds (14) where k= 8 — pyi(t~ po) Using the state and co-state equations of systems (1) and (°) ‘we simplify the Eq. (14) and finally obtain £() ae (aw = PPS(RE (as ~ 1) — dda) + (kPSE + dp8)(A5 ~ Aa). +A ASE + byQ — d5)(25 ~)P ~ pes IMIC) + {ERS(A1 ~ Aa) + Ekigin +0 +d +a) ~ (byha + 04+ 09s )RE — BER — ha) = UAE + RC — 2 RSE ~ (by 40-4 d+ EEN pS ~ (ADSKE + pS) + dpSKE Ag + (KE(0g ~ 31) ~ ddA ~KSE +10 — d5)P +(A~RE(A—ISE + 5,Q - 5) ~ KSUSE — bE ~oE — dE ab) + bu(baE — (by +€-+d)Q) — ACA —KSE + 610 —d)]pC2s ~21) +A ASE 4 byQ ~ dS) (dis ~ (KE 4 d)Ry + KERS)P + BSU?2s + pdas(A— KSE + by — dS) (15) ‘The above equation can be writen in the form ps KEp(As — br) +dp?S125— 1) @ (aH and then we can solve the singular control as 110 = [kpEOs ~ pPS(KECa2 — 91) ~ di) + (kpSE + dS) (is — + (A= KSE + byQ —AdS\(25 ~ 1) p— pSdas| and 2(0) = (RESCH ~ fa) + Ekin(ba +0 +440) = (aha + ha + 025 )KE — RE? (Aa — ha) — dk E +h ~ Ba RSE (ba +d-+0 +0 E))pS — (dpkSE + d?pS)Ay + dpkSEA2 + (RE(Ag — 41) — dh (A — KSE + b,Q — dS)p +1A~k(A~KSE + b,Q — dS)E —KS(ISE— Eo — dE aE) + bytbab —(b; +€-44)Q) ~ (A= ASE +b, —d5)]pC2s 21) 6 ‘M. Monat, Jaa and SK Nona cto hans, Sotons and Fatal 195 2020) 10889 F(AMKSE + b1Q —d5)(dis — (KE+d) 2 + KEA,)P + DSPs + pag(A— KSE + b,Q — dS) Moreover in order to satisfy the Generalized Legendre-clesh Condition for the singular contol t© be optimal, we require A Si(B) = 2100) to be negative [25]. Therefore we summarize fhe Control profile on a nontrivial interval inthe following way: a <0, then MC) em an oS it So, then Mee) =a, aH oe) iT Fy =O then Manin) = — Ft Hence the control is optimal provided y(t) < 0 and a ~ tip =o. 5. Numerical simulation We study numerical results in two different cases, frst for fixed control and second when the control has been applied optimally. First, we consider the values of parameters in Table 1, for numeri- cal simulations. Since 5 is the disease induced mortality rate and is the natural death rate, hence 8 > d, Using these parameters and the initial conditions as S(O) = 500, £(0) = 10, Q(0) =5, (0) 1, R(0) =0, we solve numerically our proposed model. (1) Fig, 2 verifies the numerical result when Ry < 1 and in this case the solutions of the model (1) converge to the DFE 13 Q Fig. 3. Dynamical behavior around FF, ‘he trisoia bration lag depicts the change of stably a Ry ~ 1 and the corresponding solutions of the madel (1) converge 0 the EE, £,(S*,E*,Q°,1",R) = (48.16, 1.21, 15.73, 384, 16633), ‘These results are discussed in the Theorem 32 and Theorem 3.4 Further, in the Theorem 3.3, we have shown that the system (1) ‘may undergo through 2 transcritical bifurcation at the threshold parametric condition Rp = 1. Therefore in Fig. 4, we demonstrate the scenarios when Ry =1 and it has been observed that the ‘model (1) possesses a transeritical bifurcation there. We now perform the sensitivity analysis to determine the higher and lower impact of some parameters on the basic repro- duction number. Ro and hence impact of such parameters on the dynamics of the proposed model (1). To perform the sensitivity analysis [22], we caleulate the normalized forward sensitivity in- {ex of Ry and these indices measures the relative change in a Ro with respect to the relative change in its parameters (Table 1). Definition 5.1. The normalized forward sensitivity index of a func- tion, Rx. #2. Ae} fOr (1 ones? os23769 090075 Estimated _ ~ i Osr6i%9 —_Do2si61 Oot4s2 Estimated bes ig. 9. Vaation o the adjoin variables when canta apptied optima J Ey] ig. 10. Vaan ofthe contol strategy of cont parameter. most populated country, has also been affected severely due to the global pandemic COVID-19. In this paper, we analyze the ef- fect of COVID-19 cases in three states in three different parts of India. These states are the Western state Maharashtra (where the ‘commercial capital of India, Mumbai situated), the northern state Delhi (where the capital of India, New Delhi is located), and the southern state Tamil Nadu (where another highly populated mega- city, Chennai is situated) and the some of rotal populations of these three states crossing two hundred millions. We apply the model system (1) to study the pandemic situation due to COVID-19 in these three states of India from 2nd March, 2020 to 27th April, 2020 and predict the future behavior of the disease in a short term basis. Although the first case of COVID-19 in India was confirmed in the state Kerala on 30th January 2020, the frst cases of COVID- 19 were reported in the state of Delhi, Tamil Nadu, and Maharash- tra respectively on 2nd, 7th, and 9th March 2020, The active, re- covered, death and confirmed COVID-19 cases of these three states are collected from the official websites of the oficial updates of coronavirus, COVID-19 in India, Government of India 36], Govern- meat of Delhi [11], Government of Maharashtca [10] and Govern- Ina values of population inthe states of Mahara, Dethl and Tami ad, Tatil Vales Maharashra Deni Tail acs Reerences 50) TS TSSS 9ATOGO 10.0143) £0) au 2a 3a Estimated 0) 98s as no Estimated 1 ° a 0. (roan.t436), Ro ° a oa r436}, E 00 Fig Ave COVID-19 cases in Nahas, ‘ment of Tamil Nadu[ 4] and presented during this period in the ‘Table 3, ‘We fit the proposed model (1) to the daily active infected and confirmed (cumulative) infected COVID-19 cases in those three states of India using the set of parameters as given in Table 4 and the intial size of the population from the Table 5. To fit these real data, we use the software Mathematica and then predict the be- havior of COVID-19 for those three states on a short term basis. In Figs. 11-Fig. 12, we respectively present the active COVID-19 cases in Maharashtra, Delhi, and Tamil Nadu for 91 days starting from 2nd March, 2020, till the 31st May 2020. Also, in Figs. 14, 15 and in Fig. 16, we present the cumulative confirmed (ie. the sum of active cases, recovered and death) COVID-19 cases of Maharashtra, Delhi, and Tamil Nadu, respectively, for the same period 7. Discussion and conclusion After world war Il, the globe has never undergone like the present scenatios that arrived due to the pandemic COVID-19. Starting from Wuhan city in December 2019, the virus SARS-CoV= 2 has spread throughout as many as 210 countries and territories and continues to increase its pandemic nature. Inthe absence of proper vaccination and treatment of COVID-19, one would rely on 0 -M. Mondat, Jana and SK Noa cto Chas, Sotons and Fractal 1952020) 10889 Fi. 12, Active COMID-9 cases in Dei Fig 18, Active COVID-I9 cases in Tat Na qualitative control of the disease rather than complete eradication Since these types of situations are unknown to present society, in these circumstances, various governments have taken some poli- cies €© control the destructive nature of the disease as much as possible. To study the dynamics of the disease, in this article, we have proposed and analyzed a classical SEIR type mathematical ‘model to incorporate the COVID-19 scenarios in the system (1) [A detalled analysis shows that the proposed system posses two equilibria, namely one disease-free and one endemic whose ex- istence and asymptotic stability criteria depend on the numerical Value of basic reproduction number Ro. We have also established that the proposed system (1) undergoes a transcritical bifurcation at the threshold Ry =. To study the impact of the government Control measures (0 prevent the extensive transmission of COVID- 19, we have introduced a control parameter M, Further, to reduce the infected individuals as well as to minimize the cost of im- plementing government control measures. an objective functional has been formed and solved using Pontryagin’s maximum princi- ple. The optimal control follows a combination of bang-bang and singular control during its application. We also have discussed the sensitivity index ofthe threshold parameter fp and found the most sensitive parameter. which has a positive impact on Rp is the dis- ease transmission rate, Next, we consider three cases of populations, namely (i) active «cases of COVID-18, (ii) confirmed cases of COVID-18, and (ii) re- covered cases of COVID-19 in three states of India, namely, western sate Maharashtra, northern state Delhi, and southern state Tamil Nadu where there are more than Gvo hundred people residing. In ‘Table 3, we provide the data of COVID-19 confirmed, active, recov- ered, and death cases of these three states stating from 2nd March ’ 3 : Fig. 1. Contrmed COMD-19 eases in Manassa Fig. 5. Coofemed COMD-9 eases in De 2020, till 27th April 2020, In Tables 4 and 5, we estimate the para- ‘metric Values associated with the model system (1). Finally, using the software Mathematica, we try to fit our model (1) to estimate the nature of COVID-19 in those three states of India. In Figs. 1 13, we present the active COVID-19 cases in Maharashtra, Delhi, and Tamil Nadu, respectively, for 91 days starting from 2nd March 2020, wll the 31st of May 2020. In particular, we use the exist ing data for the first 57 days (ull 27th April 2020) collecting from the websites of the Government of India and WHO to forecast the probable active COVID-19 cases in those three states for the rest of 34 days, It should be noted that here we can forecast the na- ture of the COVID-19 for the short term only as the Governmental policy would change in time to time resulting in the correspond- Ing changes in the associated parameters of the proposed model system. From those three figures, we can claim that on 31st May 2020, the active cases of the three states Maharashtra, Delhi, and ‘Tamil Nadu, will be around 46183, 3344, and 411 respectively. Sim- ilaly, in Figs. 14, 15 and in Fig. 16, we respectively present the cu- ‘mulative confirmed (ie. the sum of active, recovered and) death) COVID-19 cases of Maharashtra, Delhi, and) Tamil Nadu starting from 2nd March 2020, tll the 31st May 2020. Further, using the existing data for the first 57 days (til 27th April 2020), the cases for the next 34 days has been predicted. It has been forecasting from those three figures that till 31st May 2020, the cumulative confirmed instances of those three states Maharashtra, Delhi, and ‘Tamil Nadu will be around 83958, 19667, and 5133 respectively. Moreover, from above figures (Fiz. 11 to 16), we can estimate the ‘number of active cases and number of cumulative confirmed eases of those three states at any day of May, 2020, The primary finding of this article is that we have derived ‘4 mathematical model that can be used to study the qualitative 1M Mando Jona and SK Nan! ak/Chas Soto and Frac 1362020) 108880 0 eb 8 Fig. 16, Confirmed COMD-19 eases in Tail Nadu ‘dynamics of COVID-19. The basic reproduction number and its sen- sitivity analysis would determine the controling procedure of the disease. Also, we have incorporated the governmental policy in ‘our mathematical model and proposed a linear objective functional considering that governmental policy as the time-dependent con- trol parameter. On the other hand, it should be noted that the es- timated values of COVID-19 in the three different states of India in three different parts are forecasted using existing parametric space ‘only. In the case ofthe first two states, Maharashtra and Delhi we see that the graph shows an increasing trend whereas for the state Tamil Nadu, the predicted data shows that the disease can be con- trolled with the existing parametric space, The main two burdens behind the control of COVID-19 are (i) the unconsciousness of the people about the disease and (ji) the high population density ex- Posed the infection when the common people are out for their es- ‘ential commodities. Moreover, the weakness of forecasting cumu- lative COVID-19 cases in the three states of India is that here the Forecast is done on otally some existing parametric conditions. But human behaviour is the most uncertain phenomena. Hence. if the corresponding parametric space has been altered then there may be some changes in the graphs of COVID-19 cases. Therefore, in this article, we are targeting to forecast the COVID-19 cases on a short term basis and henceforth there would be very litte chance to change in corresponding parametric space. But the framework of the current model gives some important insights into the dynam ies of COVID-19 spread and control, Further, our simulation works ‘suggest that both the quarantine and governmental intervention strategies like lockéown, media coverage on social distancing and public hygiene can play an important role in diminishing COVID- 19 transmission, However, the successes of these strategies mainly rely on the proper implementation of the process. In our future work, we need to extend our model by incorporating some essen tial biological as well as epidemiological factors. To be given some ‘examples, we may mention the acquired immunity in humans, dif- ferential susceptibility and infectivity of humans to COVID-19 in- fection, and spatial heterogeneity Declaration of Competing Interest ‘The authors declare that they have no known competing finan- ial interests or personal relationships that could have appeared to Influence the work reported in this paper. (CRediT authorship contribution statement Manotosh Mandal: Conceptualization, Methodology, Software Formal analysis, Investigation, Validation, Writing - original ‘raft. Soovoojeet Jana: Conceptualization, Validation, Methodol- ‘ogy, Writing - original draft, Writing - review & editing. Visual- ization, Swapan Kumar Nandi: Validation, Formal analysis, Inves- tigation. Anupam Khatua: Investigation, Resources, Data curation Sayani Adak: Resources, Data curation, TK. Kar: Conceptualization, ‘Writing - review & editing, Visualization, Supervision, ‘Acknowledgment “The work of, Jana and S. Adak are partially supported by Dept of Science and Technology & Biotechnology, Govt. of West Ben- gal (vide memo no, 201 (Sanc.)ST)PISRT/16G-12/2018 dated 19- (02-2019). Research of A. Khatua is financially supported by De- partment of Science and Technology-INSPIRE, Government of India (No. DST/INSPIRE Felloship/2016/IF1G0667, dated: 21st Septem- ber, 2016), Further the authors are very much grateful to the ‘anonymous reviewers and Prof. Gabriel Mindlin, Handling Editor ‘of the journal, for their constructive comments and useful sugges- tions, which have helped us significant improvement of the article References 10] BermoulD, Esa dune nowwelle analyse de ts morale cause pat Ls pte ero: Ales Math iy Acad Ray Se Pas 1760, p 05. (2) kat Cota GL. Ora difereal equations Boston Gin; 1082 1B] Buono B. Onto A LacignolaD. Cabal siabity of an SR epidemic ‘model mith formation dcpendent vaccination Math Bost 200816818. (4) Charabrty Gosh Ream forests ane sk assessment of nove como ‘ivr (COMIG-0) cases a datardeven analy, Chive Stone Fractals 202 4101016 cnaos 202010850. {51 Dekmann 0 JA), JAPH On the defetion on the computation o the nase reproduction number ratio yin models for infectious dass i Reeroge- ‘as population. J Math Bol 1090:28:365-82 (6) Dresscne PV, Watmough J Rerecucton numbers and sub-thresolé endemic quia for compartment mades of dice tansmisson. Math Bas 2hna:i0-29-48 10) Fano 0, Piazza F. Analysis and forecast of COMID-19 spreading In China, Katy nd France Chaos Sotons Fractals 20213410076. (a) Fleming Wat Rise! RW. Deterministic and stochatle optimal conto. New Yorke Springer Veg: 175 [0] Fafa Gk, bers Nc IAP H. The wetapopuatin dynamics ofan infections

    You might also like