1. bookAHEAD OF PRINT
Zeitschriftendaten
License
Format
Zeitschrift
eISSN
2444-8656
Erstveröffentlichung
01 Jan 2016
Erscheinungsweise
2 Hefte pro Jahr
Sprachen
Englisch
access type Uneingeschränkter Zugang

Design of Morlet wavelet neural network to solve the non-linear influenza disease system

Online veröffentlicht: 17 Jan 2022
Volumen & Heft: AHEAD OF PRINT
Seitenbereich: -
Eingereicht: 17 Jun 2021
Akzeptiert: 24 Sep 2021
Zeitschriftendaten
License
Format
Zeitschrift
eISSN
2444-8656
Erstveröffentlichung
01 Jan 2016
Erscheinungsweise
2 Hefte pro Jahr
Sprachen
Englisch
Abstract

In this study, the solution of the non-linear influenza disease system (NIDS) is presented using the Morlet wavelet neural networks (MWNNs) together with the optimisation procedures of the hybrid process of global/local search approaches. The genetic algorithm (GA) and sequential quadratic programming (SQP), that is, GA-SQP, are executed as the global and local search techniques. The mathematical form of the NIDS depends upon four groups: susceptible S(y), infected I(y), recovered R(y) and cross-immune individuals C(y). To solve the NIDS, an error function is designed using NIDS and its leading initial conditions (ICs). This error function is optimised with a combination of MWNNs and GA-SQP to solve for all the groups of NIDS. The comparison of the obtained solutions and Runge–Kutta results is presented to authenticate the correctness of the designed MWNNs along with the GA-SQP for solving NIDS. Moreover, the statistical operators using different measures are presented to check the reliability and constancy of the MWNNs along with the GA-SQP to solve the NIDS.

Introduction

Influenza is a serious disease caused by viruses that affect the lungs, the upper breathing organs, throat, bronchi and nose. Its recovery rate is very high, but only with proper medical care. A high infection rate is noticed in older adults or in those with severe kidney, heart and lung problems, diabetes or cancer. The epidemic rate of influenza is estimated to be between 5% and 15% per year of the population, caused by upper respiratory tract disease. Worldwide, annual epidemics are seen with 3–5 million illness cases and the number of deaths is estimated at around 250,000–500,000 [1]. In mathematical form, epidemiological systems are demonstrated by the ordinary autonomous non-linear differential systems with assumptions of the time-independent parameters. Such biological models involve the variable state to the infected, recovered, susceptible and transmitted vectors.

A number of schemes have been tested to solve the non-linear influenza disease system (NIDS). A few of them are as follows. Astuti et al. [2] suggested a differential transformation step-by-step scheme to solve the influenza virus disease-resistant system. Alzahrani et al. [3] proposed a numerical approach for solving a fractional influenza pandemic system. Erdem et al. [4] discussed the influenza SIQR model using the imperfect quarantine system. Sun et al. [5] stated an optimisation based multi-objective system to allocate patients during an influenza epidemic. González-Parra et al. [6] designed a fractional epidemiologic system based on simulating an influenza A epidemic. Ghanbari et al. [7] worked towards examining the two systems of avian influenza epidemics related to the derivatives of fractal–fractional with memorabilia and power from Mittag–Leffler. Tchuenche et al. [8] attempted to enhance the media coverage impacts on the dynamics of human influenza communications. Schulze-Horsel et al. [9] worked on the dynamics of infection as well as virus-induced apoptosis using the influenza vaccine production in cellular philosophy. Hovav et al. [10] discussed the system of network flow to manage the inventory and allocating of influenza vaccines in a healthcare supply chain management. Patel et al. [11] applied genetic algorithms (GAs) for optimal vaccination strategies for the pandemic influenza system. Kanyiri et al. [12] worked on the applications of optimal control for influenza along with the antiviral resistance and pulmonary congestion.

NIDS is divided into four groups: (i) susceptible S(y), (ii) infectious I(y), (iii) recovered R(y) and (iv) cross-immune C(y). The mathematical form of NIDS along with the initial conditions (ICs) is given as follows [13]: {S(y)=μ(μ+βI(y))S(y)+γC(y),S(0)=u1,I(y)=β(σC(y)+S(y))I(u)(α+μ)I(y),I(0)=u2,R(y)=αI(y)+β(1σ)I(y)C(y)(δ+μ)R(y),R(0)=u3,C(y)=δR(y)(μ+βI(y)+γ)C(y),C(0)=u4, \left\{\begin{aligned} S^{\prime}(y) &=\mu-(\mu+\beta I(y)) S(y)+\gamma C(y), & S(0) &=u_{1}, \\ I^{\prime}(y) &=\beta(\sigma C(y)+S(y)) I(u)-(\alpha+\mu) I(y), & I(0) &=u_{2}, \\ R^{\prime}(y) &=\alpha I(y)+\beta(1-\sigma) I(y) C(y)-(\delta+\mu) R(y), & R(0) &=u_{3}, \\ C^{\prime}(y) &=\delta R(y)-(\mu+\beta I(y)+\gamma) C(y), & C(0) &=u_{4}, \end{aligned}\right.

where, u1, u2, u3 and u4 are the ICs, β represents the rate of transmission from the individuals of S(y) to I(y), σ is the exposed individuals based on cross-immune that are shifted in a unit time to the communicable subpopulation [14]. Moreover, the infectious, infected and cross-immune are defined as δ−1, γ−1 and α−1, respectively.

The purpose of the current investigation is to treat NIDS numerically by using the Morlet wavelet neural networks (MWNNs) together with the optimisation procedures in the hybrid process of global/local search approaches. The GA and sequential quadratic programming (SQP), that is, GA-SQP, are executed as global and local search techniques. There are various applications in which stochastic computing approaches have been applied, such as COVID-19-based SITR dynamics [15, 16], singular fractional models [17, 18], prey-predator model [19], delay singular functional model [20, 21], dengue fever model [22], higher order non-linear singular systems [2325], non-linear mosquito release system in heterogeneous atmosphere [26] and multi-singular differential systems [27, 28]. Keeping in view these recognised submissions, the authors are motivated to solve NIDS using the MWNNS and GA-SQP. Some motivational factors of MWNNs using the GA-SQP are briefly as follows:

The proposed MWNNs using GA-SQP provides impressive numerical solutions of NIDS.

Steady, reliable and stable numerical outcomes of NIDS authenticate the worth of the proposed form of MWNNs using GA-SQP.

The values of the absolute error (AE) show the best performances, which demonstrate the consistency of the proposed MWNNs using GA-SQP.

The numerical performance of the scheme is certified using different statistical annotations to solve the NIDS for multiple independent runs.

The proposed MWNNs using GA-SQP is smoothly executed to solve NIDS with inclusive, easy-to-understand and smooth operations.

The remainder of the paper is categorised as follows: Section 2 depicts the designed MWNNs using the GA-SQP methodology along with statistical procedures. Section 3 provides the results and simulation. Section 4 deals with the final comments and future research directions.

Methodology: MWNNs–GA-SQP

The structure of the MWNNs–GA-SQP is presented in this section based on two phases for solving NIDS:

A merit function is proposed based on MWNNs using the GA-SQP to solve NIDS.

Some major settings are provided to improve the merit function using the methodology of GA-SQP.

Structure of MWNNs

The mathematical design to solve NIDS is divided into four groups: susceptible (S), infectious (I), recovered (R) and cross-immune (C). The performance through the proposed results of these groups are Ŝ,Î,R̂\hat{S}, \hat{I}, \hat{R} and Ĉ, written as: =[i=1mrS,iL(wS,iy+sS,i),i=1mrI,iL(wI,iy+sI,i),i=1mSR,iL(wR,iy+sR,i),i=1mrC,iL(wC,iy+sC,i)],[S^(y),I^(y),R^(y),C^(y)]=[i=1mrS,iL(wS,iy+sS,i),i=1mrI,iL(wI,iy+sI,i),i=1mrR,iL(wR,iy+sR,i),i=1mrC,iL(wC,iy+sC,i)], \begin{aligned} [\hat{S}(y), \hat{I}(y), \hat{R}(y), \hat{C}(y)] &=\left[\begin{array}{ll} \sum_{i=1}^{m} r_{S, i} L\left(w_{S, i} y+s_{S, i}\right), & \sum_{i=1}^{m} r_{I, i} L\left(w_{I, i} y+s_{I, i}\right), \\ \sum_{i=1}^{m} S_{R, i} L\left(w_{R, i} y+s_{R, i}\right), & \sum_{i=1}^{m} r_{C, i} L\left(w_{C, i} y+s_{C, i}\right) \end{array}\right], \\ \left[\hat{S}^{\prime}(y), \hat{I}^{\prime}(y), \hat{R}^{\prime}(y), \hat{C}^{\prime}(y)\right] &=\left[\begin{array}{ll} \sum_{i=1}^{m} r_{S, i} L^{\prime}\left(w_{S, i} y+s_{S, i}\right), & \sum_{i=1}^{m} r_{I, i} L^{\prime}\left(w_{I, i} y+s_{I, i}\right), \\ \sum_{i=1}^{m} r_{R, i} L^{\prime}\left(w_{R, i} y+s_{R, i}\right), & \sum_{i=1}^{m} r_{C, i} L^{\prime}\left(w_{C, i} y+s_{C, i}\right) \end{array}\right], \end{aligned}

W is the unknown weight vector, given as:

W = [WS, WI, WR, WC], for WS = [rS,ωS, sS], WI = [rI, ωI, sI], WR = [rR, ωR, sR] and WC = [aC, ωC, bC], where rS=[rS,1,rS,2,,rS,m],rI=[rI,1,rI,2,,rI,m],rR=[rR,1,rR,2,,rR,m],rC=[rC,1,rC,2,,rC,m],wS=[wS,1,wS,2,,wS,m],wI=[wI,1,wI,2,,wI,m],wR=[wR,1,wR,2,,wR,m],wC=[wC,1,wC,2,,wC,m],sS=[sS,1,sS,2,,sS,m],sI=[sI,1,sI,2,,sI,m],sR=[sR,1,sR,2,,sR,m],sC=[sC,1,sC,2,,sC,m]. \begin{aligned} \boldsymbol{r}_{S} &=\left[r_{S, 1}, r_{S, 2}, \ldots, r_{S, m}\right], & \boldsymbol{r}_{I} &=\left[r_{I, 1}, r_{I, 2}, \ldots, r_{I, m}\right], & \boldsymbol{r}_{R} &=\left[r_{R, 1}, r_{R, 2}, \ldots, r_{R, m}\right], \\ \boldsymbol{r}_{C} &=\left[r_{C, 1}, r_{C, 2}, \ldots, r_{C, m}\right], & \boldsymbol{w}_{S} &=\left[w_{S, 1}, w_{S, 2}, \ldots, w_{S, m}\right], & \boldsymbol{w}_{I} &=\left[w_{I, 1}, w_{I, 2}, \ldots, w_{I, m}\right], \\ \boldsymbol{w}_{R} &=\left[w_{R, 1}, w_{R, 2}, \ldots, w_{R, m}\right], & \boldsymbol{w}_{C} &=\left[w_{C, 1}, w_{C, 2}, \ldots, w_{C, m}\right], & \boldsymbol{s}_{S} &=\left[s_{S, 1}, s_{S, 2}, \ldots, s_{S, m}\right], \\ \boldsymbol{s}_{I} &=\left[s_{I, 1}, s_{I, 2}, \ldots, s_{I, m}\right], & \boldsymbol{s}_{R} &=\left[s_{R, 1}, s_{R, 2}, \ldots, s_{R, m}\right], & \boldsymbol{s}_{C} &=\left[s_{C, 1}, s_{C, 2}, \ldots, s_{C, m}\right] . \end{aligned}

The updated form of NIDS using the Morlet function L(y) = (1 + exp(−y))−1 [2931] is given as: [S^(y),I^(y),R^(y),C^(y)]=[i=1mrS,icos(1.75(wS,iy+sS,i))e0.5(wS,yy+ss,i)2,i=1mrI,icos(1.75(wI,iy+sI,i))e0.5(wl,y+sl,i)2,i=1mrR,icos(1.75(wR,iy+sR,i))e0.5(wR,jy+sR,i)2,i=1mrC,icos(1.75(wC,iy+sC,i))e0.5(wC,iy+sC,i)2],[S^(y),I^(y),R^(y),C^(y)]=ddy[i=1mrS,icos(1.75(wS,iy+sS,i))e0.5(wS,iy+ss,i)2,i=1mrI,icos(1.75(wI,iy+sI,i))e0.5(wI,jy+sl,i)2,i=1mrR,icos(1.75(wR,iy+sR,i))e0.5(wR,iy+sR,i)2,i=1mrC,icos(1.75(wC,iy+sC,i))e0.5(wC,iy+sC,i)2]. \begin{gathered} {[\hat{S}(y), \hat{I}(y), \hat{R}(y), \hat{C}(y)]=\left[\begin{array}{l} \sum_{i=1}^{m} r_{S, i} \cos \left(1.75\left(w_{S, i} y+s_{S, i}\right)\right) e^{-0.5\left(w_{S, y} y+s_{s, i}\right)^{2}}, \\ \sum_{i=1}^{m} r_{I, i} \cos \left(1.75\left(w_{I, i} y+s_{I, i}\right)\right) e^{-0.5\left(w_{l, y}+s_{l, i}\right)^{2}}, \\ \sum_{i=1}^{m} r_{R, i} \cos \left(1.75\left(w_{R, i} y+s_{R, i}\right)\right) e^{-0.5\left(w_{R, j} y+s_{R, i}\right)^{2}}, \\ \sum_{i=1}^{m} r_{C, i} \cos \left(1.75\left(w_{C, i} y+s_{C, i}\right)\right) e^{-0.5\left(w_{C, i} y+s_{C, i}\right)^{2}} \end{array}\right],} \\ {\left[\hat{S}^{\prime}(y), \hat{I}^{\prime}(y), \hat{R}^{\prime}(y), \hat{C}^{\prime}(y)\right]=\frac{d}{d y}\left[\begin{array}{l} \sum_{i=1}^{m} r_{S, i} \cos \left(1.75\left(w_{S, i} y+s_{S, i}\right)\right) e^{-0.5\left(w_{S, i} y+s_{s, i}\right)^{2}}, \\ \sum_{i=1}^{m} r_{I, i} \cos \left(1.75\left(w_{I, i} y+s_{I, i}\right)\right) e^{-0.5\left(w_{I, j} y+s_{l, i}\right)^{2}}, \\ \sum_{i=1}^{m} r_{R, i} \cos \left(1.75\left(w_{R, i} y+s_{R, i}\right)\right) e^{-0.5\left(w_{R, i} y+s_{R, i}\right)^{2}}, \\ \sum_{i=1}^{m} r_{C, i} \cos \left(1.75\left(w_{C, i} y+s_{C, i}\right)\right) e^{-0.5\left(w_{C, i} y+s_{C, i}\right)^{2}} \end{array}\right] .} \end{gathered}

A merit function is given as: Ξ=Ξ1+Ξ2+Ξ3+Ξ4+Ξ5, \Xi=\Xi_{1}+\Xi_{2}+\Xi_{3}+\Xi_{4}+\Xi_{5}, Ξ1=1Ni=1N[r̂iμ+(βÎi+μ)ŜiγĈi]2, \Xi_{1}=\frac{1}{N} \sum_{i=1}^{N}\left[\hat{r}_{i}^{\prime}-\mu+\left(\beta \hat{I}_{i}+\mu\right) \hat{S}_{i}-\gamma \hat{C}_{i}\right]^{2}, Ξ2=1Ni=1N[ÎiβŜiÎi+μÎi+σÎiαR̂i]2, \Xi_{2}=\frac{1}{N} \sum_{i=1}^{N}\left[\hat{I}_{i}^{\prime}-\beta \hat{S}_{i}^{\prime} \hat{I}_{i}^{\prime}+\mu \hat{I}_{i}^{\prime}+\sigma \hat{I}_{i}^{\prime}-\alpha \hat{R}_{i}\right]^{2}, Ξ3=1Ni=1N[(R̂)i+αÎiβ(1σ)ĈiÎi+(μ+σ)R̂i]2, \Xi_{3}=\frac{1}{N} \sum_{i=1}^{N}\left[\left(\hat{R}^{\prime}\right)_{i}+\alpha \hat{I}_{i}^{\prime}-\beta(1-\sigma) \hat{C}_{i}^{\prime} \hat{I}_{i}^{\prime}+(\mu+\sigma) \hat{R}_{i}^{\prime}\right]^{2}, Ξ4=1Ni=1N[ĈiδR̂i+(γ+βÎi+μ)Ĉi]2, \Xi_{4}=\frac{1}{N} \sum_{i=1}^{N}\left[\hat{C}^{\prime}{ }_{i}-\delta \hat{R}_{i}^{\prime}+\left(\gamma+\beta \hat{I}_{i}^{\prime}+\mu\right) \hat{C}_{i}^{\prime}\right]^{2}, Ξ4=1Ni=1N[ĈiδR̂i+(γ+βÎi+μ)Ĉi]2, \Xi_{5}=\frac{1}{4}\left[\left(\hat{S}_{0}^{\prime}-u_{1}\right)^{2}+\left(\hat{I}_{0}^{\prime}-u_{2}\right)^{2}+\left(\hat{R}_{0}^{\prime}-u_{3}\right)^{2}++\left(\hat{C}_{0}^{\prime}-u_{4}\right)^{2}\right],

where Ŝi=S(yi),Îi=I(yi),R̂i=R(yi)\hat{S}_{i}=S\left(y_{i}\right), \hat{I}_{i}=I\left(y_{i}\right), \hat{R}_{i}=R\left(y_{i}\right) and Ĉi = C (µi). In Systems 5–8, Ξ1, Ξ2, Ξ3 and Ξ4 indicate the merit functions based on System (1), whereas Ξ5 represents the merit function based on ICs.

Optimisation measures: MWNNs–GA-SQP

This section provides a detailed procedure of the designed MWNNs together with GA-SQP for solving NIDS. The designed MWNNs structure using GA-SQP for solving NIDS is shown in Figure 1.

Fig. 1

Proposed framework of MWNNs using the GA-SQP to solve the biological-based NIDS. GA, genetic algorithm; MWNNs, Morlet wavelet neural networks; SQP, sequential quadratic programming.

GA was first applied by professor John Holland in 1975 [21] to present a simple representation of natural selection. GA grows with the population of applicant results. A genetic-based search initiates with a random (initial) population, then operators, as crossover, selection and mutation. It is applied one after another to get a new chromosome generation in which the projected excellence over all the chromosomes is improved over that of the preceding generation. This procedure is repeated till the termination standard is encountered, and the best values of the chromosomes of the final generation are described as the terminal solution. The evolutionary algorithms based on GA are broadly applied by researchers due to their capability of controlling the effectiveness, robustness, divergence-free, not to become fixed in local minima, consistent and efficient as compared with other mathematical heuristic solvers. Recently, GA is being applied in the network anomaly detection system [32], wellhead back pressure control system [33], optimising bank lending decisions [34], green vehicle routing systems [35], adaptive anomaly-based intrusion detection system [36], population initialisation with dispatching rules [37], heat conduction system [38], path planning in a dynamic field [39], Thomas–Fermi system [40] and non-linear HIV infection system [41].

SQP is one of the efficient, local search, speedy and rapid optimisation scheme generally applied to solve constrained/unconstrained systems. SQP is executed in numerous optimisation models of numerous complexes as well as non-stiff systems. Presently, it is used to investigate the guidewire deformation in blood vessels [42], in the power system stabiliser design [43], optimal control of rapid cooperative rendezvous [44], 3D deformable prostate model pose estimation in minimally invasive surgery [45], deterministic constrained production optimisation of hydrocarbon reservoirs [46], prediction differential system [47] and in the optimisation of an auxetic jounce bumper [48]. To switch the sluggishness of GA, hybridisation of the GA-SQP process is implemented along with the necessary steps, as provided in Table 1.

Optimisation through MWNNs–GA-SQP to solve NIDS.

Process of GA starts
  Inputs: The number of chromosome are selected as: W = [u, W, b]
  Population: Set of chromosomes are given as:
  W = [WS, WI, WR, WC], for WS = [rS, ωS, sS], WI = [rI, ωI, sI], WR = [rR, ωR, sR] and WC = [rC, ωC ,sC].
  Output: Global weights (GA) vectors are WB.GA
  Initialisation: To select the variables, adjust the WB.GA.
  Evaluation of FIT: Modify FIT (Ξ) in the population (P) for Eqs 49.
  Stopping process: Stop if [Ξ = 10−21], [Generations = 100], [TolFun = 10−21]. [StallLimit = 130], [TolCon = 10−19] & [PopSize = 180] achieved.
  Go to [storage]
  Ranking: Rank specific W in the selected population for Ξ.
  Storage: Save WB.GA, Ξ, iterations, function counts and time.
  End of GA
SQP process starts
  Inputs: WB.GA.
  Output: The best weight values of the GA-SQP are designated as WGA.SQP.
  Initialise: Assignments, WB.GA, iterations and other performances.
  Stopping standards: Terminate if one can achieve [Ξ = 10−18], [TolFun = 10−21], [Iterations = 500], [TolCon = TolX = 10−18] & [MaxFunEvals = 180000].
  FIT Assessment: Compute Ξ and W for Eqs 49.
  Adjustments: Regulate ‘fmincon’ for SQP and Ξ for Eqs 49.
  Accumulate: Transform WGA.SQP, time, iterations, Ξ and function counts.
  SQP End

GA, genetic algorithm; Max, maximum; MWNNs, Morlet wavelet neural networks; NIDS, non-linear influenza disease system; SQP, sequential quadratic programming.

Performance operators

The mathematical notations using the statistical-based operators, including ‘mean absolute deviation (MAD)’, ‘variance account for (VAF)’, ‘Theil’s inequality coefficient (TIC)’ and ‘semi interquartile (S.I.R)’, along with their global performances accessible to solve the biological based NIDS, are written as: {[[VAFS,VAFI,VAFR,VAFC]=[(1var(SkS^k)var(Sk))×100(1var(Ikf^k)var(Sk))×100(1var(Ikl^k)var(Ik))×100(1var(RkR^k)var(Rk))×100][EVAFS,EVAFp,EVAFR,EVAFR]=[100VAFS,100VAFI,100VAFR,100VAFC]]. \left\{\left[\begin{array}{l} {\left[\operatorname{VAF}_{S}, \operatorname{VAF}_{I}, \operatorname{VAF}_{R}, \operatorname{VAF}_{C}\right]=\left[\begin{array}{l} \left(1-\frac{\operatorname{var}\left(S_{k}-\hat{S}_{k}\right)}{\operatorname{var}\left(S_{k}\right)}\right) \times 100 \\ \left(1-\frac{\operatorname{var}\left(I_{k}-\hat{f}_{k}\right)}{\operatorname{var}\left(S_{k}\right)}\right) \times 100 \\ \left(1-\frac{\operatorname{var}\left(I_{k}-\hat{l}_{k}\right)}{\operatorname{var}\left(I_{k}\right)}\right) \times 100 \\ \left(1-\frac{\operatorname{var}\left(R_{k}-\hat{R}_{k}\right)}{\operatorname{var}\left(R_{k}\right)}\right) \times 100 \end{array}\right]} \\ {\left[\mathrm{EVAF}_{S}, \mathrm{EVAF}_{p}, \mathrm{EVAF}_{R}, \mathrm{EVAF}_{\mathrm{R}}\right]=\left[\begin{array}{l} 100-\mathrm{VAF}_{S}, 100-\mathrm{VAF}_{I}, \\ 100-\mathrm{VAF}_{R}, 100-\mathrm{VAF}_{C} \end{array}\right]} \end{array}\right] .\right. { S.I.R =12(Q1Q3)Q1=1st  quartile & Q3=3rd  quartile,  \left\{\begin{array}{l} \text { S.I.R }=-\frac{1}{2}\left(Q_{1}-Q_{3}\right) \\ Q_{1}=1^{\text {st }} \text { quartile & } Q_{3}=3^{\text {rd }} \text { quartile, } \end{array}\right. [TICS,TICI,TICR,TICc]=[1nk=1n(SkS^k)2(1nk=1nSk2+1nk=1nSk2,1nk=1n(IkIk)2(1nk=1nIk2+1nk=1nIk2,1nk=1n(RkR^k)2(1nk=1nRk2+1nk=1nRk2),1nk=1n(CkC^k)2(1nk=1nCi2+1nk=1nC^k2)], \left[\mathrm{TIC}_{S}, \mathrm{TIC}_{I}, \mathrm{TIC}_{R}, \mathrm{TIC}_{c}\right]=\left[\begin{array}{c} \frac{\sqrt{\frac{1}{n} \sum_{k=1}^{n}\left(S_{k}-\hat{S}_{k}\right)^{2}}}{\left(\sqrt{\frac{1}{n} \sum_{k=1}^{n} S_{k}^{2}}+\sqrt{\frac{1}{n} \sum_{k=1}^{n} S_{k}^{2}}\right.}, \frac{\sqrt{\frac{1}{n} \sum_{k=1}^{n}\left(I_{k}-I_{k}\right)^{2}}}{\left(\sqrt{\frac{1}{n} \sum_{k=1}^{n} I_{k}^{2}+\sqrt{\frac{1}{n} \sum_{k=1}^{n} I_{k}^{2}}}\right.}, \\ \frac{\sqrt{\frac{1}{n} \sum_{k=1}^{n}\left(R_{k}-\hat{R}_{k}\right)^{2}}}{\left(\sqrt{\frac{1}{n} \sum_{k=1}^{n} R_{k}^{2}}+\sqrt{\frac{1}{n} \sum_{k=1}^{n} R_{k}^{2}}\right)}, \frac{\sqrt{\frac{1}{n} \sum_{k=1}^{n}\left(C_{k}-\hat{C}_{k}\right)^{2}}{\left(\sqrt{\frac{1}{n} \sum_{k=1}^{n} C_{i}^{2}}+\sqrt{\frac{1}{n} \sum_{k=1}^{n} \hat{C}_{k}^{2}}\right)}}{} \end{array}\right], [MADS,MADI,MADR,MADC]=[k=1n|SkŜk|,k=1n|IkÎk|,k=1n|RkR̂k|,k=1n|CkĈk|] \left[\mathrm{MAD}_{S}, \operatorname{MAD}_{I}, \operatorname{MAD}_{R}, \operatorname{MAD}_{C}\right]=\left[\begin{array}{cc} \sum_{k=1}^{n}\left|S_{k}-\hat{S}_{k}\right|, & \sum_{k=1}^{n}\left|I_{k}-\hat{I}_{k}\right|, \\ \sum_{k=1}^{n}\left|R_{k}-\hat{R}_{k}\right|, & \sum_{k=1}^{n}\left|C_{k}-\hat{C}_{k}\right| \end{array}\right]

where, Ŝ,Î,R̂\hat{S}, \hat{I}, \hat{R} and Ĉ represent the approximate forms of the solutions.

Simulations and numerical results

The comparative presentations of the obtained numerical results and the Runge–Kutta solutions are specialised to check the accuracy of the MWNNs–GA-SQP. Furthermore, the statistical representations are specified to check the precision, accuracy and reliability of the proposed scheme. The efficient form of NIDS using the suitable parameters is accessible as: {S(y)=0.02(50I(y)+0.02)S(y)+0.5C(y),S(0)=0.8I(y)=50(S(y)+0.05C(y))I(y)73.02I(y),I(0)=0.1R(y)=73I(y)+47.5C(y)I(y)1.02R(y),R(0)=0.04C(y)=R(y)(0.52+50I(y))C(y),C(0)=0.06 \left\{\begin{aligned} S^{\prime}(y) &=0.02-(50 I(y)+0.02) S(y)+0.5 C(y), & & S(0)=0.8 \\ I^{\prime}(y) &=50(S(y)+0.05 C(y)) I(y)-73.02 I(y), & & I(0)=0.1 \\ R^{\prime}(y) &=73 I(y)+47.5 C(y) I(y)-1.02 R(y), & & R(0)=0.04 \\ C^{\prime}(y) &=R(y)-(0.52+50 I(y)) C(y), & & C(0)=0.06 \end{aligned}\right.

A merit function based on NIDS (14) is written as: Ξ=1Nk=1N([Ŝk0.02+50ŜkÎk+0.02Ŝk0.5Ĉk]2+[Îk50ŜkÎk2.5ĈkÎk+73.02Îk]2+[R̂k73Îk47.5ĈkÎk+1.02R̂k]2+[ĈkR̂k+0.52Ĉk50ÎkĈk]2)+14[(Ŝ00.8)2+(Î00.1)2+(R̂00.04)2+(Ĉ00.06)2]. \begin{aligned} \Xi=& \frac{1}{N} \sum_{k=1}^{N}\left(\begin{array}{l} {\left[\hat{S}_{k}^{\prime}-0.02+50 \hat{S}_{k} \hat{I}_{k}+0.02 \hat{S}_{k}-0.5 \hat{C}_{k}\right]^{2}+\left[\hat{I}_{k}^{\prime}-50 \hat{S}_{k} \hat{I}_{k}-2.5 \hat{C}_{k} \hat{I}_{k}+73.02 \hat{I}_{k}\right]^{2}} \\ +\left[\hat{R}_{k}^{\prime}-73 \hat{I}_{k}-47.5 \hat{C}_{k} \hat{I}_{k}+1.02 \hat{R}_{k}\right]^{2}+\left[\hat{C}^{\prime}{ }_{k}-\hat{R}_{k}+0.52 \hat{C}_{k}-50 \hat{I}_{k} \hat{C}_{k}\right]^{2} \end{array}\right) \\ &+\frac{1}{4}\left[\left(\hat{S}_{0}-0.8\right)^{2}+\left(\hat{I}_{0}-0.1\right)^{2}+\left(\hat{R}_{0}-0.04\right)^{2}+\left(\hat{C}_{0}-0.06\right)^{2}\right] .\end{aligned}

The demonstration of the performance is presented to solve NIDS using the designed MWNNs–GA-SQP for multiple trials along with 30 variations. The obtained solutions of NIDS are stated in the form of best weight vector values, which are presented in Eqs 1619. The graphical representations of these best weight vector values are illustrated in Figure 2. Ŝ(Ξ)=17.16cos(1.75(0.0090Ξ+0.1085))e0.5(0.0090Ξ+0.1085)2+0.383cos(1.75(17.533Ξ3.0806))e0.5(17.53Ξ3.0806)2+1.2417cos(1.75(8.6442Ξ+4.1544))e0.5(8.6442Ξ+4.1544)23.7922cos(1.75(11.40Ξ14.176))e0.5(11.40Ξ14.176)2+3.4256cos(1.75(0.2127Ξ2.2218))e0.5(0.2127Ξ2.2218)2, \begin{aligned} \hat{S}(\Xi)=&-17.16 \cos (1.75(0.0090 \Xi+0.1085)) e^{-0.5(0.0090 \Xi+0.1085)^{2}} \\ &+0.383 \cos (1.75(-17.533 \Xi-3.0806)) e^{-0.5(-17.53 \Xi-3.0806)^{2}} \\ &+1.2417 \cos (1.75(8.6442 \Xi+4.1544)) e^{-0.5(8.6442 \Xi+4.1544)^{2}} \\ &-3.7922 \cos (1.75(-11.40 \Xi-14.176)) e^{-0.5(-11.40 \Xi-14.176)^{2}} \\ &+3.4256 \cos (1.75(0.2127 \Xi-2.2218)) e^{-0.5(0.2127 \Xi-2.2218)^{2}}, \end{aligned} Î(Ξ)=17.1643cos(1.75(20Ξ2.5219))e0.5(20.00Ξ2.5219)2+0.3836cos(1.75(0.0461Ξ0.6859))e0.5(0.04619Ξ0.6859)2+1.2417cos(1.75(1.4619Ξ+0.8549))e0.5(1.4619Ξ+0.8549)23.4256cos(1.75(9.906Ξ2.1412))e0.5(0.1054Ξ+0.4437)2+3.4256cos(1.75(2.3213Ξ6.8886))e0.5(0.1424Ξ+0.518)2, \begin{aligned} \hat{I}(\Xi)=&-17.1643 \cos (1.75(-20 \Xi-2.5219)) e^{-0.5(-20.00 \Xi-2.5219)^{2}} \\ &+0.3836 \cos (1.75(0.0461 \Xi-0.6859)) e^{-0.5(0.04619 \Xi-0.6859)^{2}} \\ &+1.2417 \cos (1.75(1.4619 \Xi+0.8549)) e^{-0.5(1.4619 \Xi+0.8549)^{2}} \\ &-3.4256 \cos (1.75(-9.906 \Xi-2.1412)) e^{-0.5(0.1054 \Xi+0.4437)^{2}} \\ &+3.4256 \cos (1.75(2.3213 \Xi-6.8886)) e^{-0.5(-0.1424 \Xi+0.518)^{2}}, \end{aligned} R̂(Ξ)=1.0285cos(1.75(16.848Ξ+1.1773))e0.5(16.848Ξ+1.1773)2+1.7527cos(1.75(0.3206Ξ+17.2099))e0.5(0.3206Ξ+17.2099)2+1.4105cos(1.75(17.2958Ξ+1.5052))e0.5(17.2958Ξ+1.5052)21.6372cos(1.75(0.07120Ξ+2.3310))e0.5(0.07120Ξ+2.3310)2+0.6731cos(1.75(0.663Ξ12.5775))e0.5(0.663Ξ12.5775)2, \begin{aligned} \hat{R}(\Xi)=&-1.0285 \cos (1.75(16.848 \Xi+1.1773)) e^{-0.5(16.848 \Xi+1.1773)^{2}} \\ &+1.7527 \cos (1.75(0.3206 \Xi+17.2099)) e^{-0.5(0.3206 \Xi+17.2099)^{2}} \\ &+1.4105 \cos (1.75(17.2958 \Xi+1.5052)) e^{-0.5(17.2958 \Xi+1.5052)^{2}} \\ &-1.6372 \cos (1.75(0.07120 \Xi+2.3310)) e^{-0.5(0.07120 \Xi+2.3310)^{2}} \\ &+0.6731 \cos (1.75(-0.663 \Xi-12.5775)) e^{-0.5(-0.663 \Xi-12.5775)^{2}}, \end{aligned} Ĉ(Ξ)=1.6283cos(1.75(0.3292Ξ2.1560))e0.5(0.3292Ξ2.1560)26.6645cos(1.75(0.0350Ξ0.4678))e0.5(0.035Ξ0.4678)22.9774cos(1.75(0.0916Ξ+2.47260))e0.5(0.0916Ξ+2.47260)20.4609cos(1.75(4.359Ξ+13.907))e0.5(4.3590Ξ+13.9070)20.1666cos(1.75(1.53431Ξ+2.5169))e0.5(1.53431Ξ+2.5169)2, \begin{aligned} \hat{C}(\Xi)=&-1.6283 \cos (1.75(0.3292 \Xi-2.1560)) e^{-0.5(0.3292 \Xi-2.1560)^{2}} \\ &-6.6645 \cos (1.75(-0.0350 \Xi-0.4678)) e^{-0.5(-0.035 \Xi-0.4678)^{2}} \\ &-2.9774 \cos (1.75(0.0916 \Xi+2.47260)) e^{-0.5(0.0916 \Xi+2.47260)^{2}} \\ &-0.4609 \cos (1.75(-4.359 \Xi+13.907)) e^{-0.5(-4.3590 \Xi+13.9070)^{2}} \\ &-0.1666 \cos (1.75(1.53431 \Xi+2.5169)) e^{-0.5(1.53431 \Xi+2.5169)^{2}}, \end{aligned}

Fig. 2

Best weight vectors set and best/mean results comparison with reference solutions to solve NIDS. NIDS, non-linear influenza disease system.

The obtained outputs are calculated using Eqs 1619 within the range of 0–1, to indicate the numerical outcomes for each group of NIDS. Figure 2(a–d) illustrates the weight vector plots based on the best solutions to solve NIDS. A comparison of the best and mean solutions using MWNNs–GA-SQP with the reference Runge–Kutta results are presented in Figure 2(e–h) for solving NIDS. It is indicated in these plots that the obtained results through MWNNs–GA-SQP overlapped with the reference solutions for each group of NIDS. This overlapping of the results indicates the excellence and precision of the designed MWNNs–GA-SQP. The AE plots for each group of NIDS are provided in Figure 3. One can find that the best values of the AE for the groups susceptible S(y), infected I(y), recovered R(y) and cross-immune C(y) based on NIDS is found to be around 10−02–10−03, 10−02–10−04, 10−04–10−05 and 10−02–10−03, respectively, whereas the mean values of the AE for the groups S(y), I(y), R(y) and C(y) are found to be around 10−01–10−02, 10−01–10−03, 10−01–10−04 and 10−01–10−03, respectively. Figure 4 signifies the convergence performances in terms of EVAF, MAD and TIC statistical operators for solving each group of NIDS. It is observed that the convergence performances of the best EVAF, MAD and TIC statistical operator of the S(y) group lie around 10−02–10−03, 10−01–10−02 and 10−05– 10−06, respectively. The best values of the I(y) group lie around 10−02–10−03, 10−03–10−04 and 10−05–10−06 for the operators EVAF, MAD and TIC, respectively. The best values of the R(y) group lie around 10−01–10−02 for the operators EVAF and MAD, while for the TIC operator, these values lie around 10−04–10−05. The best values of the C(y) group lie around 10−02–10−03, 10−01–10−03 and 10−05–10−06 for the operators EVAF, MAD and TIC, respectively. These encouraging suggestions confirm the correctness of the MWNNs–GA-SQP for solving each group of NIDS.

Fig. 3

Best and mean values of the AE for each group of NIDS. AE, absolute error; NIDS, non-linear influenza disease system.

Fig. 4

Statistical performances based on EVAF, MAD and TIC operators for solving NIDS. MAD, mean absolute deviation; NIDS, non-linear influenza disease system; TIC, Theil’s inequality coefficient.

The graphical plots based on the statistical measures to authorise the performance of convergence are given in Figure 5 to solve NIDS. The performance of the TIC operator using multiple executions to solve NIDS. It is observed the maximum (Max) number of trials based on the S(y), I(y), R(y) and C(y) groups of the TIC operators are found around 10−04–10−06, 10−05–10−06, 10−03–10−06 and 10−05–10−06, respectively. The performances of the EVAF values for the S(y), I(y), R(y) and C(y) groups lie around 10−05–10−07. The performances of the EVAF values for the S(y), I(y), R(y) and C(y) groups lie around 10−01–10−04. These best values of the executions based on the MWNNs–GA-SQP are found suitable based on the TIC, EVAF and MAD operators for solving NIDS.

Fig. 5

Convergence performances based on TIC, EVAF and RMSE operators for solving NIDS. NIDS, non-linear influenza disease system.

The statistical representations are provided in Tables 25, based on the operators Minimum (Min), standard deviation (STD), Mean, Max, Median (Med) and S.I.R to authenticate the precision and accurateness for solving each group of NIDS. The Max values for the S(y), I(y), R(y) and C(y) groups based on NIDS lie around 10−02–10−04, 10−01–10−02, 10−02–10−03 and 10−01–10−02. The Min values for the S(y), I(y), R(y) and C(y) groups based on NIDS lie around 10−03–10−06, 10−03–10−05, 10−03–10−05 and 10−04–10−07. The Med values for the S(y), I(y), R(y) and C(y) groups based on NIDS lie around 10−02–10−05, 10−02–10−04, 10−03–10−04 and 10−01–10−04. The Mean, S.I.R and STD values for S(y), I(y), R(y) and C(y) groups based on NIDS lie around 10−02–10−03, 10−02–10−04 and 10−03–10−04. These small calculated values designate the worth and performance of MWNNs–GA-SQP to solve each group of NIDS. One can find through these attained measures that the proposed MWNNs–GA-SQP is stable, precise and accurate.

Statistical performances for NIDS-based S(y).

y S(y)
  Max Min Med Mean S.I.R STD
0 9.97407E−02 6.23657E−06 9.67647E−02 3.15861E−03 6.56937E−04 1.11176E−03
0.1 3.82797E−03 2.46073E−03 2.63760E−03 3.28652E−02 3.74939E−04 1.50172E−01
0.2 1.59218E−03 2.35321E−03 2.86086E−04 3.36978E−02 4.41197E−03 1.41380E−01
0.3 7.52762E−04 2.33924E−03 1.43482E−04 3.35508E−02 7.04131E−03 1.27429E−01
0.4 1.17681E−03 1.74469E−03 1.49107E−04 3.20838E−02 8.72803E−03 1.14515E−01
0.5 8.45225E−04 1.43288E−03 8.54966E−05 3.07057E−02 9.01342E−03 1.02999E−01
0.6 9.64376E−04 6.38340E−04 6.69333E−05 2.97428E−02 9.36168E−03 9.27501E−02
0.7 1.03468E−03 1.08089E−03 8.11832E−05 2.88456E−02 9.76687E−03 8.35818E−02
0.8 1.04732E−03 9.68113E−04 6.50947E−05 2.78143E−02 1.02683E−02 7.53055E−02
0.9 9.93449E−04 6.87790E−04 6.98467E−05 2.65912E−02 1.08763E−02 6.77636E−02
1 8.66289E−04 3.87491E−04 6.32949E−05 2.51101E−02 1.15130E−02 6.08310E−02

Max, maximum; Med, median; Min, minimum; NIDS, non-linear influenza disease system; S.I.R, semi interquartile.

Statistical performances for NIDS-based I(y).

y I(y)
  Min Max Med Mean S.I.R SD
0 1.07524E−02 1.13718E−05 6.43473E−04 1.84408E−03 1.81226E−02 1.12267E−03
0.1 1.02042E−01 1.55703E−04 9.97639E−02 2.50909E−03 2.90364E−04 7.08815E−03
0.2 1.03305E−01 2.36329E−03 1.01710E−01 7.67239E−03 9.27342E−05 2.05399E−02
0.3 1.01679E−01 2.59357E−04 1.00308E−01 1.30883E−02 1.08303E−04 3.23014E−02
0.4 1.00068E−01 1.83367E−04 9.69585E−02 1.72975E−02 6.27781E−05 4.25479E−02
0.5 9.73448E−02 1.67571E−03 9.31885E−02 2.05823E−02 5.26592E−05 5.12053E−02
0.6 9.38826E−02 1.71384E−03 9.00952E−02 2.32492E−02 5.88914E−05 5.83514E−02
0.7 8.99105E−02 1.69484E−03 8.65303E−02 2.55060E−02 5.75666E−05 6.41060E−02
0.8 8.55744E−02 1.52268E−03 8.22485E−02 2.74916E−02 5.39135E−05 6.85933E−02
0.9 8.09763E−02 1.14571E−03 7.80244E−02 2.93020E−02 4.83546E−05 7.19385E−02
1 7.61938E−02 5.69917E−04 7.35171E−02 3.10092E−02 3.62539E−05 7.42639E−02

Max, maximum; Med, median; Min, minimum; NIDS, non-linear influenza disease system; S.I.R, semi interquartile.

Statistical performances for NIDS-based R(y).

y R(y)
  Max Min Med Mean S.I.R STD
0 6.92243E−03 5.64028E−05 2.11655E−04 3.15861E−03 9.26468E−04 7.81009E−02
0.1 1.51990E−02 8.69332E−03 7.09041E−03 3.28652E−02 1.78168E−02 2.56713E−03
0.2 2.92093E−02 5.01355E−03 2.44201E−02 3.36978E−02 1.51010E−02 3.86235E−04
0.3 4.14911E−02 5.07207E−03 3.94198E−02 3.35508E−02 1.34338E−02 1.77153E−04
0.4 5.40406E−02 4.99083E−03 5.19292E−02 3.20838E−02 1.28378E−02 2.18649E−04
0.5 6.43792E−02 4.44331E−03 6.24911E−02 3.07057E−02 1.23623E−02 1.43347E−04
0.6 7.31617E−02 4.24610E−03 7.10555E−02 2.97428E−02 1.20395E−02 1.42089E−04
0.7 8.03111E−02 4.63968E−03 7.80977E−02 2.88456E−02 1.17625E−02 1.55918E−04
0.8 8.60489E−02 5.44280E−03 8.35771E−02 2.78143E−02 1.16541E−02 1.16323E−04
0.9 9.05767E−02 6.39784E−03 8.76171E−02 2.65912E−02 1.14688E−02 1.19646E−04
1 9.40773E−02 7.34102E−03 9.11587E−02 2.51101E−02 1.11275E−02 1.06769E−04

Max, maximum; Med, median; Min, minimum; NIDS, non-linear influenza disease system; S.I.R, semi interquartile.

Statistical performances for NIDS-based C(y).

y C(y)
  Max Min Med Mean S.I.R STD
0 1.48512E−02 2.50654E−03 1.85049E−04 1.84408E−03 3.85021E−04 2.18319E−03
0.1 1.90381E−01 2.03580E−04 1.86341E−01 2.50909E−03 3.27146E−02 8.05712E−02
0.2 1.77459E−01 1.05623E−04 1.74888E−01 7.67239E−03 2.46695E−02 8.27583E−02
0.3 1.60388E−01 7.96272E−06 1.58166E−01 1.30883E−02 2.07562E−02 8.11572E−02
0.4 1.46077E−01 5.88240E−06 1.41058E−01 1.72975E−02 1.89140E−02 7.90228E−02
0.5 1.32592E−01 5.20223E−07 1.26230E−01 2.05823E−02 1.73278E−02 7.64063E−02
0.6 1.19829E−01 3.96760E−07 1.14065E−01 2.32492E−02 1.59848E−02 7.32956E−02
0.7 1.08056E−01 1.44197E−06 1.03949E−01 2.55060E−02 1.49896E−02 6.99375E−02
0.8 9.73313E−02 1.37818E−06 9.40407E−02 2.74916E−02 1.37403E−02 6.64160E−02
0.9 8.76205E−02 1.07879E−06 8.53022E−02 2.93020E−02 1.25955E−02 6.27128E−02
1 7.88533E−02 3.73973E−07 7.71953E−02 3.10092E−02 1.09303E−02 5.88248E−02

Max, maximum; Med, median; Min, minimum; NIDS, non-linear influenza disease system; S.I.R, semi interquartile.

The global operators of [G-TIC], [G-MAD] and [G-EVAF] for multiple independent trials to solve NIDS using the proposed MWNNs–GA-SQP are tabulated in Table 6. These global performances based on Mean lie around 10−02–10−03, 10−06–10−07 and 10−01–10−02, whereas the global S.I.R performances are found around 10−02 to 10−03, 10−07–10−08 and 10−02–10−03 for each group of NIDS. These ideal close performances attained through global measures show the precision, accuracy and correctness of the designed MWNNs–GA-SQP for solving all the groups of NIDS.

Global operators based on TIC, MAD and EVAF values to solve NIDS.

Index (G-MAD) (G-TIC) (G-EVAF)
  Mean S.I.R Mean S.I.R Mean S.I.R
S(y) 6.66623E−02 1.60694E−03 2.89392E−06 3.44759E−08 7.20734E−02 2.11643E−02
I(y) 7.47583E−03 1.14516E−02 9.71475E−07 7.52206E−07 7.09781E−01 2.08050E−03
R(y) 9.25308E−02 7.73955E−03 4.16727E−06 2.24806E−07 6.77949E−02 2.90708E−03
C(y) 4.47326E−02 1.63317E−02 2.11608E−06 5.07713E−07 8.00158E−01 2.75884E−02

MAD, mean absolute deviation; NIDS, non-linear influenza disease system; S.I.R, semi interquartile.

Conclusions

This current work is related to solve NIDS by exploiting MWNNs using the optimisation procedures of the hybrid process of global/local search approaches. The GA as a global approach and SQP as a local search approach have been implemented as an optimisation procedure to solve the non-linear biological model. The influenza model is based on four groups: susceptible S(y), infected I(y), recovered R(y) and cross-immune C(y); the numerical results of each group of the influenza system are presented using the proposed MWNNs–GA-SQP. The Morlet wavelet function is exploited the first time to solve the non-linear influenza system using 15 numbers of variables together with the optimisation procedures of GA-SQP. For the exactness of the MWNNs–GA-SQP, a comparison of the proposed results with the reference solutions is performed. The overlapping of the results around 4–5 decimal places is noticed as a result of this comparison and this accuracy develops in the consistency of the designed MWNNs–GA-SQP. The particular performances of the MWNNs–GA-SQP through the statistical RMSE, TIC, EVAF operators are observed for 40 trials to solve NIDS. These statistical operator performances are accomplished based on the higher precision level to solve NIDS. Moreover, statistics by Max, Min, STD, S.I.R, Med and Mean were further validated to solve NIDS using the proposed MWNNs–GA-SQP. The global performances of the operators are found to be in good measure through the Mean and S.I.R operators, which prove the validity and correctness of the MWNNs–GA-SQP for solving NIDS.

In future, the proposed MWNNs–GA-SQP can be used to solve the systems of higher order models, fluids problems and non-linear biological systems [4958].

Fig. 1

Proposed framework of MWNNs using the GA-SQP to solve the biological-based NIDS. GA, genetic algorithm; MWNNs, Morlet wavelet neural networks; SQP, sequential quadratic programming.
Proposed framework of MWNNs using the GA-SQP to solve the biological-based NIDS. GA, genetic algorithm; MWNNs, Morlet wavelet neural networks; SQP, sequential quadratic programming.

Fig. 2

Best weight vectors set and best/mean results comparison with reference solutions to solve NIDS. NIDS, non-linear influenza disease system.
Best weight vectors set and best/mean results comparison with reference solutions to solve NIDS. NIDS, non-linear influenza disease system.

Fig. 3

Best and mean values of the AE for each group of NIDS. AE, absolute error; NIDS, non-linear influenza disease system.
Best and mean values of the AE for each group of NIDS. AE, absolute error; NIDS, non-linear influenza disease system.

Fig. 4

Statistical performances based on EVAF, MAD and TIC operators for solving NIDS. MAD, mean absolute deviation; NIDS, non-linear influenza disease system; TIC, Theil’s inequality coefficient.
Statistical performances based on EVAF, MAD and TIC operators for solving NIDS. MAD, mean absolute deviation; NIDS, non-linear influenza disease system; TIC, Theil’s inequality coefficient.

Fig. 5

Convergence performances based on TIC, EVAF and RMSE operators for solving NIDS. NIDS, non-linear influenza disease system.
Convergence performances based on TIC, EVAF and RMSE operators for solving NIDS. NIDS, non-linear influenza disease system.

Statistical performances for NIDS-based I(y).

y I(y)
  Min Max Med Mean S.I.R SD
0 1.07524E−02 1.13718E−05 6.43473E−04 1.84408E−03 1.81226E−02 1.12267E−03
0.1 1.02042E−01 1.55703E−04 9.97639E−02 2.50909E−03 2.90364E−04 7.08815E−03
0.2 1.03305E−01 2.36329E−03 1.01710E−01 7.67239E−03 9.27342E−05 2.05399E−02
0.3 1.01679E−01 2.59357E−04 1.00308E−01 1.30883E−02 1.08303E−04 3.23014E−02
0.4 1.00068E−01 1.83367E−04 9.69585E−02 1.72975E−02 6.27781E−05 4.25479E−02
0.5 9.73448E−02 1.67571E−03 9.31885E−02 2.05823E−02 5.26592E−05 5.12053E−02
0.6 9.38826E−02 1.71384E−03 9.00952E−02 2.32492E−02 5.88914E−05 5.83514E−02
0.7 8.99105E−02 1.69484E−03 8.65303E−02 2.55060E−02 5.75666E−05 6.41060E−02
0.8 8.55744E−02 1.52268E−03 8.22485E−02 2.74916E−02 5.39135E−05 6.85933E−02
0.9 8.09763E−02 1.14571E−03 7.80244E−02 2.93020E−02 4.83546E−05 7.19385E−02
1 7.61938E−02 5.69917E−04 7.35171E−02 3.10092E−02 3.62539E−05 7.42639E−02

Statistical performances for NIDS-based C(y).

y C(y)
  Max Min Med Mean S.I.R STD
0 1.48512E−02 2.50654E−03 1.85049E−04 1.84408E−03 3.85021E−04 2.18319E−03
0.1 1.90381E−01 2.03580E−04 1.86341E−01 2.50909E−03 3.27146E−02 8.05712E−02
0.2 1.77459E−01 1.05623E−04 1.74888E−01 7.67239E−03 2.46695E−02 8.27583E−02
0.3 1.60388E−01 7.96272E−06 1.58166E−01 1.30883E−02 2.07562E−02 8.11572E−02
0.4 1.46077E−01 5.88240E−06 1.41058E−01 1.72975E−02 1.89140E−02 7.90228E−02
0.5 1.32592E−01 5.20223E−07 1.26230E−01 2.05823E−02 1.73278E−02 7.64063E−02
0.6 1.19829E−01 3.96760E−07 1.14065E−01 2.32492E−02 1.59848E−02 7.32956E−02
0.7 1.08056E−01 1.44197E−06 1.03949E−01 2.55060E−02 1.49896E−02 6.99375E−02
0.8 9.73313E−02 1.37818E−06 9.40407E−02 2.74916E−02 1.37403E−02 6.64160E−02
0.9 8.76205E−02 1.07879E−06 8.53022E−02 2.93020E−02 1.25955E−02 6.27128E−02
1 7.88533E−02 3.73973E−07 7.71953E−02 3.10092E−02 1.09303E−02 5.88248E−02

Statistical performances for NIDS-based S(y).

y S(y)
  Max Min Med Mean S.I.R STD
0 9.97407E−02 6.23657E−06 9.67647E−02 3.15861E−03 6.56937E−04 1.11176E−03
0.1 3.82797E−03 2.46073E−03 2.63760E−03 3.28652E−02 3.74939E−04 1.50172E−01
0.2 1.59218E−03 2.35321E−03 2.86086E−04 3.36978E−02 4.41197E−03 1.41380E−01
0.3 7.52762E−04 2.33924E−03 1.43482E−04 3.35508E−02 7.04131E−03 1.27429E−01
0.4 1.17681E−03 1.74469E−03 1.49107E−04 3.20838E−02 8.72803E−03 1.14515E−01
0.5 8.45225E−04 1.43288E−03 8.54966E−05 3.07057E−02 9.01342E−03 1.02999E−01
0.6 9.64376E−04 6.38340E−04 6.69333E−05 2.97428E−02 9.36168E−03 9.27501E−02
0.7 1.03468E−03 1.08089E−03 8.11832E−05 2.88456E−02 9.76687E−03 8.35818E−02
0.8 1.04732E−03 9.68113E−04 6.50947E−05 2.78143E−02 1.02683E−02 7.53055E−02
0.9 9.93449E−04 6.87790E−04 6.98467E−05 2.65912E−02 1.08763E−02 6.77636E−02
1 8.66289E−04 3.87491E−04 6.32949E−05 2.51101E−02 1.15130E−02 6.08310E−02

Statistical performances for NIDS-based R(y).

y R(y)
  Max Min Med Mean S.I.R STD
0 6.92243E−03 5.64028E−05 2.11655E−04 3.15861E−03 9.26468E−04 7.81009E−02
0.1 1.51990E−02 8.69332E−03 7.09041E−03 3.28652E−02 1.78168E−02 2.56713E−03
0.2 2.92093E−02 5.01355E−03 2.44201E−02 3.36978E−02 1.51010E−02 3.86235E−04
0.3 4.14911E−02 5.07207E−03 3.94198E−02 3.35508E−02 1.34338E−02 1.77153E−04
0.4 5.40406E−02 4.99083E−03 5.19292E−02 3.20838E−02 1.28378E−02 2.18649E−04
0.5 6.43792E−02 4.44331E−03 6.24911E−02 3.07057E−02 1.23623E−02 1.43347E−04
0.6 7.31617E−02 4.24610E−03 7.10555E−02 2.97428E−02 1.20395E−02 1.42089E−04
0.7 8.03111E−02 4.63968E−03 7.80977E−02 2.88456E−02 1.17625E−02 1.55918E−04
0.8 8.60489E−02 5.44280E−03 8.35771E−02 2.78143E−02 1.16541E−02 1.16323E−04
0.9 9.05767E−02 6.39784E−03 8.76171E−02 2.65912E−02 1.14688E−02 1.19646E−04
1 9.40773E−02 7.34102E−03 9.11587E−02 2.51101E−02 1.11275E−02 1.06769E−04

Global operators based on TIC, MAD and EVAF values to solve NIDS.

Index (G-MAD) (G-TIC) (G-EVAF)
  Mean S.I.R Mean S.I.R Mean S.I.R
S(y) 6.66623E−02 1.60694E−03 2.89392E−06 3.44759E−08 7.20734E−02 2.11643E−02
I(y) 7.47583E−03 1.14516E−02 9.71475E−07 7.52206E−07 7.09781E−01 2.08050E−03
R(y) 9.25308E−02 7.73955E−03 4.16727E−06 2.24806E−07 6.77949E−02 2.90708E−03
C(y) 4.47326E−02 1.63317E−02 2.11608E−06 5.07713E−07 8.00158E−01 2.75884E−02

Optimisation through MWNNs–GA-SQP to solve NIDS.

Process of GA starts
  Inputs: The number of chromosome are selected as: W = [u, W, b]
  Population: Set of chromosomes are given as:
  W = [WS, WI, WR, WC], for WS = [rS, ωS, sS], WI = [rI, ωI, sI], WR = [rR, ωR, sR] and WC = [rC, ωC ,sC].
  Output: Global weights (GA) vectors are WB.GA
  Initialisation: To select the variables, adjust the WB.GA.
  Evaluation of FIT: Modify FIT (Ξ) in the population (P) for Eqs 49.
  Stopping process: Stop if [Ξ = 10−21], [Generations = 100], [TolFun = 10−21]. [StallLimit = 130], [TolCon = 10−19] & [PopSize = 180] achieved.
  Go to [storage]
  Ranking: Rank specific W in the selected population for Ξ.
  Storage: Save WB.GA, Ξ, iterations, function counts and time.
  End of GA
SQP process starts
  Inputs: WB.GA.
  Output: The best weight values of the GA-SQP are designated as WGA.SQP.
  Initialise: Assignments, WB.GA, iterations and other performances.
  Stopping standards: Terminate if one can achieve [Ξ = 10−18], [TolFun = 10−21], [Iterations = 500], [TolCon = TolX = 10−18] & [MaxFunEvals = 180000].
  FIT Assessment: Compute Ξ and W for Eqs 49.
  Adjustments: Regulate ‘fmincon’ for SQP and Ξ for Eqs 49.
  Accumulate: Transform WGA.SQP, time, iterations, Ξ and function counts.
  SQP End

[1] World Health Organization (WHO). Influenza Overview. Available from: http://www.who.int/mediacentre/factsheets/fs211/en/ World Health Organization (WHO) Influenza Overview Available from: http://www.who.int/mediacentre/factsheets/fs211/en/ Search in Google Scholar

[2] Astuti, F. et al., 2019, June. Multi-step differential transform method for solving the influenza virus model with disease resistance. In IOP Conference Series: Materials Science and Engineering (Vol. 546, No. 5, p. 052013). IOP Publishing. Astuti F. 2019 June Multi-step differential transform method for solving the influenza virus model with disease resistance In IOP Conference Series: Materials Science and Engineering 546 5 052013 IOP Publishing 10.1088/1757-899X/546/5/052013 Search in Google Scholar

[3] Alzahrani, E.O. et al., 2020. Comparison of numerical techniques for the solution of a fractional epidemic model. The European Physical Journal Plus, 135(1), p. 110. Alzahrani E.O. 2020 Comparison of numerical techniques for the solution of a fractional epidemic model The European Physical Journal Plus 135 1 110 10.1140/epjp/s13360-020-00183-4 Search in Google Scholar

[4] Erdem, M., et al., 2017. Mathematical analysis of an SIQR influenza model with imperfect quarantine. Bulletin of Mathematical Biology, 79(7), pp. 1612-1636. Erdem M. 2017 Mathematical analysis of an SIQR influenza model with imperfect quarantine Bulletin of Mathematical Biology 79 7 1612 1636 10.1007/s11538-017-0301-628608046 Search in Google Scholar

[5] Sun, L., et al., 2014. Multi-objective optimization models for patient allocation during a pandemic influenza outbreak. Computers & Operations Research, 51, pp. 350-359. Sun L. 2014 Multi-objective optimization models for patient allocation during a pandemic influenza outbreak Computers & Operations Research 51 350 359 10.1016/j.cor.2013.12.001 Search in Google Scholar

[6] González-Parra, G., et al., 2014. (H1N1). Mathematical methods in the Applied Sciences, 37(15), pp. 2218-2226. González-Parra G. 2014 (H1N1) Mathematical methods in the Applied Sciences 37 15 2218 2226 10.1002/mma.2968 Search in Google Scholar

[7] Ghanbari, B. et al., 2019. Analysis of two avian influenza epidemic models involving fractal-fractional derivatives with power and Mittag-Leffler memories. Chaos: An Interdisciplinary Journal of Nonlinear Science, 29(12), p. 123113. Ghanbari B. 2019 Analysis of two avian influenza epidemic models involving fractal-fractional derivatives with power and Mittag-Leffler memories Chaos: An Interdisciplinary Journal of Nonlinear Science 29 12 123113 10.1063/1.511728531893661 Search in Google Scholar

[8] Tchuenche, J.M., et al., 2011. The impact of media coverage on the transmission dynamics of human influenza. BMC Public Health, 11(S1), p. S5. Tchuenche J.M. 2011 The impact of media coverage on the transmission dynamics of human influenza BMC Public Health 11 S1 S5 10.1186/1471-2458-11-S1-S5331758521356134 Search in Google Scholar

[9] Schulze-Horsel, J., et al., 2009. Infection dynamics and virus-induced apoptosis in cell culture-based influenza vaccine production—Flow cytometry and mathematical modeling. Vaccine, 27(20), pp. 2712-2722. Schulze-Horsel J. 2009 Infection dynamics and virus-induced apoptosis in cell culture-based influenza vaccine production—Flow cytometry and mathematical modeling Vaccine 27 20 2712 2722 10.1016/j.vaccine.2009.02.02719428884 Search in Google Scholar

[10] Hovav, S., et al., 2015. A network flow model for inventory management and distribution of influenza vaccines through a healthcare supply chain. Operations Research for Health Care, 5, pp. 49-62. Hovav S. 2015 A network flow model for inventory management and distribution of influenza vaccines through a healthcare supply chain Operations Research for Health Care 5 49 62 10.1016/j.orhc.2015.05.003 Search in Google Scholar

[11] Patel, R., et al., 2005. Finding optimal vaccination strategies for pandemic influenza using genetic algorithms. Journal of theoretical biology, 234(2), pp. 201-212. Patel R. 2005 Finding optimal vaccination strategies for pandemic influenza using genetic algorithms Journal of theoretical biology 234 2 201 212 10.1016/j.jtbi.2004.11.03215757679 Search in Google Scholar

[12] Kanyiri, C.W., et al., 2020. Application of optimal control to influenza pneumonia coinfection with antiviral resistance. Computational and Mathematical Methods in Medicine, 2020. Kanyiri C.W. 2020 Application of optimal control to influenza pneumonia coinfection with antiviral resistance Computational and Mathematical Methods in Medicine 2020 10.1155/2020/5984095709154832256682 Search in Google Scholar

[13] Jódar, L., et al., 2008. Nonstandard numerical methods for a mathematical model for influenza disease. Mathematics and Computers in simulation, 79(3), pp. 622-633. Jódar L. 2008 Nonstandard numerical methods for a mathematical model for influenza disease Mathematics and Computers in simulation 79 3 622 633 10.1016/j.matcom.2008.04.008 Search in Google Scholar

[14] Casagrandi, R., et al., 2006. The SIRC model and influenza A. Mathematical biosciences, 200(2), pp. 152-169. Casagrandi R. 2006 The SIRC model and influenza A Mathematical biosciences 200 2 152 169 10.1016/j.mbs.2005.12.02916504214 Search in Google Scholar

[15] Umar, M. et al., 2021. Integrated neuro-swarm heuristic with interior-point for nonlinear SITR model for dynamics of novel COVID-19. Alexandria Engineering Journal. Umar M. 2021 Integrated neuro-swarm heuristic with interior-point for nonlinear SITR model for dynamics of novel COVID-19 Alexandria Engineering Journal 10.1016/j.aej.2021.01.043 Search in Google Scholar

[16] Umar, M. et al., 2020. A Stochastic intelligent computing with neuro-evolution heuristics for nonlinear SITR system of novel COVID-19 dynamics. Symmetry, 12(10), p. 1628. Umar M. 2020 A Stochastic intelligent computing with neuro-evolution heuristics for nonlinear SITR system of novel COVID-19 dynamics Symmetry 12 10 1628 10.3390/sym12101628 Search in Google Scholar

[17] Sabir, Z., Zahoor Raja, M.A. and Baleanu, D., 2021. Fractional Mayer Neuro-swarm heuristic solver for multifractional Order doubly singular model based on Lane-Emden equation. Fractals. Fractals, p. 2040033. Sabir Z. Zahoor Raja M.A. Baleanu D. 2021 Fractional Mayer Neuro-swarm heuristic solver for multifractional Order doubly singular model based on Lane-Emden equation. Fractals. Fractals 2040033 10.1142/S0218348X2140017X Search in Google Scholar

[18] Sabir, Z. et al., 2021. A novel design of fractional Meyer wavelet neural networks with application to the nonlinear singular fractional Lane-Emden systems. Alexandria Engineering Journal, 60(2), pp. 2641-2659. Sabir Z. 2021 A novel design of fractional Meyer wavelet neural networks with application to the nonlinear singular fractional Lane-Emden systems Alexandria Engineering Journal 60 2 2641 2659 10.1016/j.aej.2021.01.004 Search in Google Scholar

[19] Umar, M. et al., 2019. Intelligent computing for numerical treatment of nonlinear prey–predator models. Applied Soft Computing, 80, pp. 506-524. Umar M. 2019 Intelligent computing for numerical treatment of nonlinear prey–predator models Applied Soft Computing 80 506 524 10.1016/j.asoc.2019.04.022 Search in Google Scholar

[20] Sabir, Z et al., 2021. Solving a novel designed second order nonlinear Lane-Emden delay differential model using the heuristic techniques. Applied Soft Computing, p. 107105. Sabir Z 2021 Solving a novel designed second order nonlinear Lane-Emden delay differential model using the heuristic techniques Applied Soft Computing 107105 10.1016/j.asoc.2021.107105 Search in Google Scholar

[21] Guirao, J. L. et al., 2020. Design and numerical solutions of a novel third-order nonlinear Emden–Fowler delay differential model. Mathematical Problems in Engineering, 2020. Guirao J. L. 2020 Design and numerical solutions of a novel third-order nonlinear Emden–Fowler delay differential model Mathematical Problems in Engineering 2020 10.1155/2020/7359242 Search in Google Scholar

[22] Umar, M. et al., 2020. A stochastic numerical computing heuristic of SIR nonlinear model based on dengue fever. Results in Physics, 19, p. 103585. Umar M. 2020 A stochastic numerical computing heuristic of SIR nonlinear model based on dengue fever Results in Physics 19 103585 10.1016/j.rinp.2020.103585 Search in Google Scholar

[23] Sabir, Z. et al., 2020. Heuristic computing technique for numerical solutions of nonlinear fourth order Emden–Fowler equation. Mathematics and Computers in Simulation. Sabir Z. 2020 Heuristic computing technique for numerical solutions of nonlinear fourth order Emden–Fowler equation Mathematics and Computers in Simulation 10.1016/j.matcom.2020.06.021 Search in Google Scholar

[24] Sabir, Z et al, 2020. Integrated intelligent computing with neuro-swarming solver for multi-singular fourth-order nonlinear Emden–Fowler equation. Computational and Applied Mathematics, 39(4), pp. 1-18. Sabir Z 2020 Integrated intelligent computing with neuro-swarming solver for multi-singular fourth-order nonlinear Emden–Fowler equation Computational and Applied Mathematics 39 4 1 18 10.1007/s40314-020-01330-4 Search in Google Scholar

[25] Sabir, Z., et al., 2020. Design of neuro-swarming-based heuristics to solve the third-order nonlinear multi-singular Emden–Fowler equation. The European Physical Journal Plus, 135(6), p. 410. Sabir Z. 2020 Design of neuro-swarming-based heuristics to solve the third-order nonlinear multi-singular Emden–Fowler equation The European Physical Journal Plus 135 6 410 10.1140/epjp/s13360-020-00424-6 Search in Google Scholar

[26] Umar, M. et al., 2020. A stochastic computational intelligent solver for numerical treatment of mosquito dispersal model in a heterogeneous environment. The European Physical Journal Plus, 135(7), pp. 1-23. Umar M. 2020 A stochastic computational intelligent solver for numerical treatment of mosquito dispersal model in a heterogeneous environment The European Physical Journal Plus 135 7 1 23 10.1140/epjp/s13360-020-00557-8 Search in Google Scholar

[27] Raja, M.A.Z. et al., 2019. Numerical solution of doubly singular nonlinear systems using neural networks-based integrated intelligent computing. Neural Computing and Applications, 31(3), pp. 793-812. Raja M.A.Z. 2019 Numerical solution of doubly singular nonlinear systems using neural networks-based integrated intelligent computing Neural Computing and Applications 31 3 793 812 10.1007/s00521-017-3110-9 Search in Google Scholar

[28] Sabir, Z. et al., 2020. FMNEICS: fractional Meyer neuro-evolution-based intelligent computing solver for doubly singular multi-fractional order Lane–Emden system. Computational and Applied Mathematics, 39(4), pp. 1-18. Sabir Z. 2020 FMNEICS: fractional Meyer neuro-evolution-based intelligent computing solver for doubly singular multi-fractional order Lane–Emden system Computational and Applied Mathematics 39 4 1 18 10.1007/s40314-020-01350-0 Search in Google Scholar

[29] Nisar, K., et al., 2021. Design of Morlet Wavelet Neural Network for Solving a Class of Singular Pantograph Nonlinear Differential Models. IEEE Access. Nisar K. 2021 Design of Morlet Wavelet Neural Network for Solving a Class of Singular Pantograph Nonlinear Differential Models IEEE Access 10.1109/ACCESS.2021.3072952 Search in Google Scholar

[30] Umar, M., et al., 2021. A novel study of Morlet neural networks to solve the nonlinear HIV infection system of latently infected cells. Results in Physics, p. 104235. Umar M. 2021 A novel study of Morlet neural networks to solve the nonlinear HIV infection system of latently infected cells Results in Physics 104235 10.1016/j.rinp.2021.104235 Search in Google Scholar

[31] Sabir, Z. et al., 2020. Novel design of Morlet wavelet neural network for solving second order Lane–Emden equation. Mathematics and Computers in Simulation, 172, pp. 1-14. Sabir Z. 2020 Novel design of Morlet wavelet neural network for solving second order Lane–Emden equation Mathematics and Computers in Simulation 172 1 14 10.1016/j.matcom.2020.01.005 Search in Google Scholar

[32] Hamamoto, A.H., et al., 2018. Network anomaly detection system using genetic algorithm and fuzzy logic. Expert Systems with Applications, 92, pp. 390-402. Hamamoto A.H. 2018 Network anomaly detection system using genetic algorithm and fuzzy logic Expert Systems with Applications 92 390 402 10.1016/j.eswa.2017.09.013 Search in Google Scholar

[33] Liang, H., et al., 2020. An improved genetic algorithm optimization fuzzy controller applied to the wellhead back pressure control system. Mechanical Systems and Signal Processing, 142, p. 106708. Liang H. 2020 An improved genetic algorithm optimization fuzzy controller applied to the wellhead back pressure control system Mechanical Systems and Signal Processing 142 106708 10.1016/j.ymssp.2020.106708 Search in Google Scholar

[34] Metawa, N., et al., 2017. Genetic algorithm-based model for optimizing bank lending decisions. Expert Systems with Applications, 80, pp. 75-82. Metawa N. 2017 Genetic algorithm-based model for optimizing bank lending decisions Expert Systems with Applications 80 75 82 10.1016/j.eswa.2017.03.021 Search in Google Scholar

[35] da Costa, P.R.D.O., et al., 2018. A genetic algorithm for a green vehicle routing problem. Electronic notes in discrete mathematics, 64, pp. 65-74. da Costa P.R.D.O. 2018 A genetic algorithm for a green vehicle routing problem Electronic notes in discrete mathematics 64 65 74 10.1016/j.endm.2018.01.008 Search in Google Scholar

[36] Resende, P.A.A., et al., 2018. Adaptive anomaly-based intrusion detection system using genetic algorithm and profiling. Security and Privacy, 1(4), p. e36. Resende P.A.A. 2018 Adaptive anomaly-based intrusion detection system using genetic algorithm and profiling Security and Privacy 1 4 e36 10.1002/spy2.36 Search in Google Scholar

[37] Vlašić, I., et al., 2019. Improving genetic algorithm performance by population initialisation with dispatching rules. Computers & Industrial Engineering, 137, p. 106030. Vlašić I. 2019 Improving genetic algorithm performance by population initialisation with dispatching rules Computers & Industrial Engineering 137 106030 10.1016/j.cie.2019.106030 Search in Google Scholar

[38] Raja, M.A.Z., et al., 2018. A new stochastic computing paradigm for the dynamics of nonlinear singular heat conduction model of the human head. The European Physical Journal Plus, 133(9), p. 364. Raja M.A.Z. 2018 A new stochastic computing paradigm for the dynamics of nonlinear singular heat conduction model of the human head The European Physical Journal Plus 133 9 364 10.1140/epjp/i2018-12153-4 Search in Google Scholar

[39] Elhoseny, M., et al., 2018. Bezier curve based path planning in a dynamic field using modified genetic algorithm. Journal of Computational Science, 25, pp. 339-350. Elhoseny M. 2018 Bezier curve based path planning in a dynamic field using modified genetic algorithm Journal of Computational Science 25 339 350 10.1016/j.jocs.2017.08.004 Search in Google Scholar

[40] Sabir, Z., et al., 2018. Neuro-heuristics for nonlinear singular Thomas-Fermi systems. Applied Soft Computing, 65, pp. 152-169. Sabir Z. 2018 Neuro-heuristics for nonlinear singular Thomas-Fermi systems Applied Soft Computing 65 152 169 10.1016/j.asoc.2018.01.009 Search in Google Scholar

[41] Umar, M., et al., 2020. Stochastic numerical technique for solving HIV infection model of CD4+ T cells. The European Physical Journal Plus, 135(6), p. 403. Umar M. 2020 Stochastic numerical technique for solving HIV infection model of CD4+ T cells The European Physical Journal Plus 135 6 403 10.1140/epjp/s13360-020-00417-5 Search in Google Scholar

[42] Li, L., Tang, et al., 2019. Investigation of guidewire deformation in blood vessels based on an SQP algorithm. Applied Sciences, 9(2), p. 280. Li L. Tang 2019 Investigation of guidewire deformation in blood vessels based on an SQP algorithm Applied Sciences 9 2 280 10.3390/app9020280 Search in Google Scholar

[43] Faraji, A., et al., 2019. A combined approach for power system stabilizer design using continuous wavelet transform and SQP algorithm. International Transactions on Electrical Energy Systems, 29(3), p. e2768. Faraji A. 2019 A combined approach for power system stabilizer design using continuous wavelet transform and SQP algorithm International Transactions on Electrical Energy Systems 29 3 e2768 10.1002/etep.2768 Search in Google Scholar

[44] Liu, G., et al., 2019. Hybrid QPSO and SQP algorithm with homotopy method for optimal control of rapid cooperative rendezvous. Journal of Aerospace Engineering, 32(4), p. 04019030. Liu G. 2019 Hybrid QPSO and SQP algorithm with homotopy method for optimal control of rapid cooperative rendezvous Journal of Aerospace Engineering 32 4 04019030 10.1061/(ASCE)AS.1943-5525.0001021 Search in Google Scholar

[45] Amparore, D., et al., 2019. Non-linear-Optimization Using SQP for 3D Deformable Prostate Model Pose Estimation in Minimally Invasive Surgery. In Science and Information Conference (pp. 477-496). Springer, Cham. Amparore D. 2019 Non-linear-Optimization Using SQP for 3D Deformable Prostate Model Pose Estimation in Minimally Invasive Surgery In Science and Information Conference 477 496 Springer Cham 10.1007/978-3-030-17795-9_35 Search in Google Scholar

[46] Liu, Z., et al., 2018. Comparison of SQP and AL algorithms for deterministic constrained production optimization of hydrocarbon reservoirs. Journal of Petroleum Science and Engineering, 171, pp. 542-557. Liu Z. 2018 Comparison of SQP and AL algorithms for deterministic constrained production optimization of hydrocarbon reservoirs Journal of Petroleum Science and Engineering 171 542 557 10.1016/j.petrol.2018.06.063 Search in Google Scholar

[47] Sabir, Z. et al., 2020. Integrated neuro-evolution heuristic with sequential quadratic programming for second-order prediction differential models. Numerical Methods for Partial Differential Equations. Sabir Z. 2020 Integrated neuro-evolution heuristic with sequential quadratic programming for second-order prediction differential models Numerical Methods for Partial Differential Equations 10.1002/num.22692 Search in Google Scholar

[48] Wang, Y., et al., 2018. Optimization of an auxetic jounce bumper based on Gaussian process metamodel and series hybrid GA-SQP algorithm. Structural and Multidisciplinary Optimization, 57(6), pp. 2515-2525. Wang Y. 2018 Optimization of an auxetic jounce bumper based on Gaussian process metamodel and series hybrid GA-SQP algorithm Structural and Multidisciplinary Optimization 57 6 2515 2525 10.1007/s00158-017-1869-z Search in Google Scholar

[49] Cordero Barbero, A., et al., 2019. Stability analysis of fourth-order iterative method for finding multiple roots of nonlinear equations. Applied Mathematics and Nonlinear Sciences, 4(1), pp. 43-56. Cordero Barbero A. 2019 Stability analysis of fourth-order iterative method for finding multiple roots of nonlinear equations Applied Mathematics and Nonlinear Sciences 4 1 43 56 10.2478/AMNS.2019.1.00005 Search in Google Scholar

[50] Sabir, Z et al., 2020. Numerical investigations to design a novel model based on the fifth order system of Emden–Fowler equations. Theoretical and Applied Mechanics Letters, 10(5), pp. 333-342. Sabir Z 2020 Numerical investigations to design a novel model based on the fifth order system of Emden–Fowler equations Theoretical and Applied Mechanics Letters 10 5 333 342 10.1016/j.taml.2020.01.049 Search in Google Scholar

[51] Guerrero Sánchez, Y., et al, 2020. Analytical and Approximate Solutions of a Novel Nervous Stomach Mathematical Model. Discrete Dynamics in Nature and Society, 2020. Guerrero Sánchez Y. 2020 Analytical and Approximate Solutions of a Novel Nervous Stomach Mathematical Model Discrete Dynamics in Nature and Society 2020 10.1155/2020/5063271 Search in Google Scholar

[52] Britton, N.F., et al., 2019. Can aphids be controlled by fungus? A mathematical model. Britton N.F. 2019 Can aphids be controlled by fungus? A mathematical model 10.2478/AMNS.2019.1.00009 Search in Google Scholar

[53] Lakshminarayana, G., et al., 2018. Peristaltic slip flow of a Bingham fluid in an inclined porous conduit with Joule heating. Applied Mathematics and Nonlinear Sciences, 3(1), pp. 41-54. Lakshminarayana G. 2018 Peristaltic slip flow of a Bingham fluid in an inclined porous conduit with Joule heating Applied Mathematics and Nonlinear Sciences 3 1 41 54 10.21042/AMNS.2018.1.00005 Search in Google Scholar

[54] Umar, M., et al, 2019. Numerical treatment for the three-dimensional Eyring-Powell fluid flow over a stretching sheet with velocity slip and activation energy. Advances in Mathematical Physics, 2019. Umar M. 2019 Numerical treatment for the three-dimensional Eyring-Powell fluid flow over a stretching sheet with velocity slip and activation energy Advances in Mathematical Physics 2019 10.1155/2019/9860471 Search in Google Scholar

[55] Brzeziński, D.W., 2018. Review of numerical methods for NumILPT with computational accuracy assessment for fractional calculus. Applied Mathematics and Nonlinear Sciences, 3(2), pp. 487-502. Brzeziński D.W. 2018 Review of numerical methods for NumILPT with computational accuracy assessment for fractional calculus Applied Mathematics and Nonlinear Sciences 3 2 487 502 10.2478/AMNS.2018.2.00038 Search in Google Scholar

[56] Sajid, T. et al, 2020. Impact of oxytactic microorganisms and variable species diffusivity on blood-gold Reiner–Philippoff nanofluid. Applied Nanoscience, pp. 1-13. Sajid T. 2020 Impact of oxytactic microorganisms and variable species diffusivity on blood-gold Reiner–Philippoff nanofluid Applied Nanoscience 1 13 10.1007/s13204-020-01581-x Search in Google Scholar

[57] Fernández-Pousa, C.R., 2018. Perfect phase-coded pulse trains generated by Talbot effect. Applied Mathematics and Nonlinear Sciences, 3(1), pp. 23-32. Fernández-Pousa C.R. 2018 Perfect phase-coded pulse trains generated by Talbot effect Applied Mathematics and Nonlinear Sciences 3 1 23 32 10.21042/AMNS.2018.1.00003 Search in Google Scholar

[58] Sajid, T et al., 2020. Impact of activation energy and temperature-dependent heat source/sink on Maxwell–Sutterby fluid. Mathematical Problems in Engineering, 2020. Sajid T 2020 Impact of activation energy and temperature-dependent heat source/sink on Maxwell–Sutterby fluid Mathematical Problems in Engineering 2020 10.1155/2020/5251804 Search in Google Scholar

Empfohlene Artikel von Trend MD

Planen Sie Ihre Fernkonferenz mit Scienceendo