IoT Technology Enabled Heuristic Model With Morlet Wavelet Neural Network for Numerical Treatment of Heterogeneous Mosquito Release Ecosystem

The utmost advancements of artificial neural networks (ANNs), software-defined networks (SDNs) and internet of things (IoT) technologies find beneficial in different applications of the smart healthcare sector. Aiming at modern technology’s use in the future development of healthcare, this paper presents an advanced heuristic based on Morlet wavelet neural network for solving the mosquito release ecosystem in a heterogeneous atmosphere. The mosquito release ecosystem is dependent of six classes, eggs density, larvae density, pupae density, mosquitoes searching for hosts density, resting mosquito’s density and mosquitoes searching for ovipositional site density. An artificial neural network with the layer structure of Morlet wavelet (MWNN) kernel is presented using the global and local search optimization schemes of genetic algorithm (GA) and active-set algorithm (ASA), i.e., MWNN-GA-ASA. The accurateness, reliability and constancy of the proposed MWNN-GA-ASA is established through comparative examinations with Adams method based numerical results to solve the proposed nonlinear system with matching of order 10−06 to 10−09. The accuracy and convergence of the proposed MWNN-GA-ASA is certified using the statistical operators based on root mean square error (RMSE), Theil’s inequality coefficient (T.I.C) and mean absolute deviation (MAD) operators.

The embedded sensing devices are employed in IoT-based systems to efficiently and economically gauge real-time environmental parameters [59], [60], [65]. A sensor is a device that can sense the change in its surrounding environment [66], [67]. The Internet of Things can fabricate and advance numerous areas of action we can discover the IoT eHealth Ecosystem [68]- [70], the loT Intelligent Transportation Ecosystem [71], the IoT Smart Home Ecosystem [72]- [74], and mosquito release Ecosystem etc as shown graphically in Fig. 2.
The release of mosquito's ecosystem is a main factor in disrupting the persistence and resurgence of numerous vector diseases. Features of spatial heterogeneity based on mosquitoes, i.e., host sites and reproduction, human association with vectors, affect the distribution and population structure of mosquitoes and the ability to control disease transmission. Mosquitoes transmit dengue, malaria, filarial, yellow fever and many other vital diseases. Malaria represents a significant spatial disparity primarily determined by climatic variations, primarily response coverage and human movement [74], [75]. At a range of 100m-1000m, the mosquito environment plays a dominant role in controlling the spread [76]. Mosquitoes as well as other animals can travel in any direction, but can travel partial distances inspired by the availability of resources. Control interfering should replicate the capacity and location of mosquitoes to move, in order to achieve a higher level of effectiveness in the collapse of the mosquito population.
The impact of vector-borne disease propagation and control was first highlighted a century ago by Ronald Ross [77]. However, he recognized that the public health community does not place a high priority on this issue. Ross stated that the density of mosquitoes depends on four variables in any region, which contain reproduction rates, mortality rates, immigration and emigration rates. Manga et al. [78] accessible that the spatial disparity in the spreading of possessions applied by mosquitoes affects their rate of dispersion and reproduction. This contributes to the variation in densities, human knowledge of vectors and the capacity to control disease communication [79], [80]. The characteristics of the resource on transport can be incredible. For example, even the presence of non-productive larval habitats can impact bite densities [81]. However, experimental investigations of mosquito dispersal are stimulating [82], [83].
Mathematical systems play a dynamic role in understanding and providing the phenomena's solutions that are stimulating for the assortment of fields, however, insufficient systems have integrated dispersal or heterogeneity wideranging characteristic of a close population vector [84]- [86]. The researchers split the mature phase of the mosquitoes into various phases [87]. To discover the effects of dispersion and heterogeneity, a system can integrate mosquito lifecycle structures, spatial heterogeneity based on mosquito properties, distribution and feeding cycle. Space systems have usually implemented the diffusion scheme which reproduces space as a constant variable. Despite the reality of dissemination models that take heterogeneity into account, it is difficult to incorporate the many factors that disturb the movement [88], [89]. For example, in areas where possessions are located in discrete patches, mosquito dispersal is more appropriately modelled using a metapopulational technique, the population is allocated into isolated spots. At each location, the population is subdivided into subgroups, resulting in a set of subgroups corresponding to different states and multiple compartmentalized systems. There are various diffusion systems have incorporated the heterogeneity present in the atmosphere on the release of disease vectors [90], [91]. Nevertheless, each has understood the aquatic phases of mosquitoes to provide a general or simple framework to model random spatial designs of mosquito control interference.
The mosquito dynamics represents a nonlinear differential system of six classes named as eggs density (E), larvae density (L), pupae density (L), mosquitoes searching based hosts density (A h ), density of resting mosquitoes (A r ) and mosquitoes searching based on ovipositional site density (A 0 ). The mathematical form of these classes based on nonlinear mosquito's dispersal system (NMDS) in the heterogeneous environment is give as [92]: The variables defined for each class of the NMDS in the heterogeneous environment (1) and the appropriate selections and ranges are given in Table 1 as reported in [92].
The motive of this work is to solve the above NMDS in the heterogeneous environment using the layer structure of Morlet wavelet (MWNN) kernel together with global and local search optimization schemes of genetic algorithm (GA) and active-set algorithm (ASA), i.e., MWNN-GA-ASA. Numerical stochastic approaches have been widely applied to solve a wide variety of applications, like delay singular functional model [93], [94], COVID-19 dynamical model [95], [96], singular fractional models [97], [98], preypredator system [99], singular nonlinear higher order models [100]- [102], HIV infection system [103], multi-singular differential systems [104], [105] and dengue fever nonlinear system [106]. Based on these renowned applications, the authors are motivated to solve the NMDS with the help of the MWNN-GA-ASA. Some main factors of the MWNN-GA-ASA are briefly discussed as: • The proposed MWNNs are designed and presented using GA-ASA optimization procedures to solve the nonlinear mosquito's dispersal system in a heterogeneous atmosphere.
• Steady, constant and trustworthy outcomes for nonlinear mosquito's dispersal system authenticate the value of the proposed MWNN-GA-ASA.
• The values of the absolute deviation from reference are found in the good agreement that further represents the dependability of the MWNN-GA-ASA.
• The MWNN-GA-ASA performance is certified using different statistics via root mean square error (RMSE), Theil's inequality coefficient (T.I.C) and mean absolute deviation (MAD) observations to solve the NMDS in a heterogeneous atmosphere for 30 independent trials.
• The proposed MWNN-GA-ASA is smoothly implemented to solve the nonlinear mosquito's dispersal system in a heterogeneous atmosphere with understandable processes, robust effective and stable. The rest of the paper is organized as: Sect 2 presents the proposed MWNN-GA-ASA and statistical procedures. Sect 3 proves the simulation of the numerical outcomes. Sect 4 indicates the final explanations and future research reports.

II. METHODOLOGY
To implement the proposed MWNN-GA-ASA, it is possible to use different IoT sensors and hardware components to detect six classes of mosquito release Ecosystem as shown in Fig. 3. The planned construction of the ANN-GA-ASA to solve NMDS in a heterogeneous atmosphere is designed in two phases: Step 1: Introduce a merit function by operating the system of MWNN.
Step 2: Necessary explanations are provided to optimize the merit function to solve NMDS in a heterogeneous atmosphere (1) by the hybrid computing GA-ASA. The proposed MWNN-GA-ASA is accessible as demonstrated in Fig. 4.

A. MODELING: MWNN-GA-ASA
The mathematical relations in case of system (1) are provided with MWNN in the proposed results form n th derivatives are given as (2), shown at the bottom of the next page, where the unknown weight vector (W) is shown as: [107]. The updated form of the system (2) is given as (3), shown at the bottom of the page.
Using the network (3), a merit function (E) is written as: where Nh = 1, The approximate solutions of eggs density (E), larvae density (L), pupae density (L), mosquitoes searching based hosts density (A h ), density of resting mosquitoes (A r ) and mosquitoes searching based on ovipositional site density (A 0 ), respectively signified asÊ m ,L m ,P m , 5 and E 6 are the merit functions associated with NDMS in a heterogeneous atmosphere and E 7 represents the initial conditions of the system (1).

B. OPTIMIZATION PROCESS: GA-ASA
In this section, a brief explanation of GA-ASA combination to optimize the merit function as shown in system (4) is provided for solving the NDMS in a heterogeneous atmosphere. Genetic algorithm is an efficient global optimization tool introduced by Holand in the last century [108]. GA is mathematical genetic procedure of humans, which is applied broadly using the optimization of decision variables in various domains. The process of GA is implemented in many applications include expenditure system of the hospitals [109], brain tumor models [110], feature collection in cancer systems [111], bismuth-borate glasses optimizations [112], prediction based differential systems [113], air blast systems of prediction [114], monorail vehicle networks [115], prediction of liver diseases [115], optimization through cloud services [117] and periodic boundary values networks [118].
Active-set approach is known as a local search process, rapidly optimize to solve the constrained/unconstrained systems generally. ASA is used to execute various stiff, complex and nonlinear systems. Recently, ASA is executed to pricing the American option [119], the actual control through optimization [120], pressure-dependent system of water distribution [121], embedded model predictive control [122], overcurrent relays in microgram optimization [123] and frictional contact models based on electrodynamic [124]. The pseudocode detail of MWNN-GA-ASA based procedures is given in Table 2, while the procedure construction is shown in Figure.4.

C. PERFORMANCE MEASURES
The performance operators to solve the NDMS in a heterogeneous atmosphere are presented using the root mean square error (RMSE) operator, mean absolute deviation (MAD) operator and Theil's inequality coefficient (TIC) operator, mathematically given as (12) and (13), shown at the bottom of the next page.

III. RESULTS AND DISCUSSION
In this section, the considerations of the results to solve the NDMS in a heterogeneous atmosphere given in system (1) are described. The relative investigations with the Adams methods precise the exactness of the proposed MWNN-GA-ASA. Moreover, statistical outcomes are plotted to authenticate the accuracy of the proposed MWNN-GA-ASA.

A. PRESENTATIONS OF NDMS IN A HETEROGENEOUS ATMOSPHERE
The efficient form of NDMS in a heterogeneous atmosphere given in system (1) using the suitable values is given as: A merit function of the model (14) is written as: The optimization of the NDMS in a heterogeneous atmosphere given in system (1) is accomplished by the hybrid based computing structure GA-ASA for 30 trials to achieve the MWNNs parameter with 15 variables of the system. The best weight values of the MWNN through GA-ASA are derived and presented graphically 3-D bar plots in Figure. 5. These weigh vectors are provided to get the estimated numerical outcomes of the system (1) for 15 number of variables. These weights are used in set of equation (3) to derive the approximate solution. Accordingly, the mathematical representations of the approximate solutions of MWNN-GA-ASA are given as (16)- (21), shown at the bottom of the page. The trained weight vectors for 15 variables based MWNN system are plotted in subfigures 5(i), 5(ii), 5(iii), 5(iv), 5(v) and 6(vi) for the classes E(x), L(x), P(x), A h (x), A r (x) and A 0 (x), respectively.The equations (16)(17)(18)(19)(20)(21) are used to show the outcomes of the NDMS in a heterogeneous atmosphere using the MWNN-GA-ASA and plot of results are given in Figures 6-10 for 15 weights or decision variable in the networks.
The graphs of AE are shown in Figure. 3. The classes E(x), L(x) and P(x) plots are given in the subfigures 6(a), while, the plots of the rest of the classes A h (x), A r (x) and A 0 (x) of the NDMS in a heterogeneous atmosphere are given in subfigures 6(b). The best AE shown in subfigure 6(a) for the classes E(x), L(x) and P(x) lie around 10 −02 to 10 −03 , 10 −03     A h (x), A r (x) and A 0 (x) of the NDMS in a heterogeneous atmosphere are illustrated in figure 8.
The best RMSE performances shown in figure 9 for the classes E(x), L(x) and P(x) lie around 10 −02 to 10 −03 , 10 −02 to 10 −04 and 10 −03 to 10 −04 , respectively. While, the best RMSE performances as presented in figure 10 for the classes A h (x), A r (x) and A 0 (x) lie about 10 −03 to 10 −04 , 10 −03 to 10 −04 and 10 −04 to 10 −05 , respectively. These accurate results, i.e., values in good agreement with the desire level for the near to perfect modelling, on different performance operator calculated for 35 trials of MWNN-GA-ASA show that most of the executions achieved higher level of accuracy for TIC and RMSE operators, which further prove the worth of the designed MWNN-GA-ASA for solving the system model.    Table 3, while these indices for P(x) VOLUME 9, 2021 and A r (x) are shown in Table 4 and the other two classes for A r (x) and A 0 (x) these metrics are tabulated in Table 5. The MIN and MAX values shows the best and worst results and a relatively small variation exist in these parameter which show the consist accuracy of MWNN-GA-ASA. The S.I.R is the difference of third and first quartiles and near to zero value of this metric is consistently achieved by MWNN-GA-ASA. The statistical performances of central tendency, i.e., mean and MED values, are found in reasonably accurate ranges for each class of the NDMS in a heterogeneous atmosphere consistently.

IV. CONCLUSION
The design of IoT technology enabled Morlet wavelet neural network is presented viably and effectively for solving a class of nonlinear mosquito's dispersal system in the heterogeneous atmosphere. A merit function is considered in accordance with the representation of differential system of mosquito's dispersal system and corresponding initial conditions with MWNNs. The optimization of merit function to solve the nonlinear biological system is performed by using the global and local search techniques, GA-ASA. One can observe that the proposed results through MWNN-GA-ASA are overlapped with the Adams results that shows the accurateness of the scheme for solving the nonlinear mosquito's dispersal system. The comparison through AE is also observed in good ranges for each class of the mosquito's dispersal system. The mosquito's dispersal system is profi- In future, the accessible MWNN-GA-ASA is promoted to solve the singular higher order, fractional models, smart cities model, and fluid dynamics systems [20], [99], [125]- [140].
ZULQURNAIN SABIR received the M.Sc. degree in mathematics from Punjab University, Lahore, Pakistan, and the M.Phil. degree in mathematics from Preston University Kohat, Islamabad Campus, Pakistan. He is currently pursuing the Ph.D. degree in mathematics from Hazara University, Mansehra, Pakistan. He has published more than 50 articles in reported international WoS journals with Impact Factors. His research interests include mathematical modeling, unsupervised neural networks, supervised neural networks, artificial intelligence, and implementation of computational techniques based on traditional and heuristic methodology. He is famous to solve singular models, functional models, fractional models, biological models and fluid models. He is a pioneer to design and solve second order pantograph Emden-Fowler model, prediction differential model, nonlinear fifth order Emden-Fowler model, nervous stomach model, and nonlinear multi-singular SITR model based on coronavirus (COVID 19 He is also working on machine and deep learning for IoT security and API security and he is also working closely with Industry. He has published over more than 200 research papers in many high impact journals and well reputed international conference proceedings in the area of computer science and computer network. His research interests include future internet (FI), information centric networks (ICN), content-centric networking (CCN), named data networking (NDN), software-defined networking (SDN), the Internet of Things (IoT), internet of everything (IoE), industrial Internet of Things (IIoT), fourth industrial revolution (IR 4.0), quantum network, information security and privacy, network/cyber security, digital forensics, applied cryptography, vehicular clouds, cloud and edge computing, and blockchain. He is a member of many professional organizations from academia and industry, including the Founding Vice-Chair, IEEE Sabah Subsection, Malaysia, a member of ACM, ACM-SIGMOBILE, ISOC, Engineers Australia, IAENG, and Park Lab. and a fellow of APAN and ITU. He is serving as an editorial board member for various high impact factor journals, including Computer Communications (Elsevier), an International journal Kybernetes (Emerald U.K.), and an International journal of Wireless Personal Communications (Springer). Additionally, he is serving as a reviewer for most of the IEEE TRANSACTIONS, IEEE ACCESS,  IEEE INTERNET INITIATIVE, INTERNET TECHNOLOGY LETTERS, Journals (Wiley,  Springer, and Elsevier). Furthermore, he is serving as a Steering Committee Member, the PC Chair, the Track Chair, a Technical Program Committee (TPC) Member of over more than 100 international conferences as IEEE GLOBECOM, IEEE R10 TENCOM, IEEE TrustCom, IEEE ICC, IEEE VTC, IEEE VNC, IEEE ICCVE, and ICCCN. He has delivered keynote talks at international conferences and universities. He is also serving as a Guest Editor for more than a dozen special issues in journals and magazines, such as IEEE, Elsevier, Springer and Wiley. He would contribute to your mission by bringing credibility to high quality of modern creative technologies research. He has extensive experience in teaching, research, and industry at key positions. He also conduct interdisciplinary research in academic and industry, integrating digital technologies in teaching at undergraduate and postgraduate levels, demonstrate his ability to convey complex information and thrive on providing a classroom experience that facilitates a high level of engagement with students of all levels. He will develop an active research program in creative technologies, and supervise postgraduate research students. He also bring international grants, local grants, collaboration with universities and industries. He has experienced and an ability to work well with students and staffs from differing academic and cultural backgrounds and at all levels.
MUHAMMAD ASIF ZAHOOR RAJA received the M.Sc. degree in mathematics from Forman Christen College Lahore, Pakistan, in 1996, the M.Sc. degree in nuclear engineering from Quaid-e-Azam, University, Islamabad, Pakistan, in 1999, and the Ph.D. degree in electronic engineering from International Islamic University, Islamabad, Pakistan, in 2011. He was involved in research and development assignment of Engineering and Scientific Commission, Pakistan, from 1999 to 2012. He is currently working as an Assistant Professor with the Department of Electrical Engineering, COMSATS Institute of Information Technology, Attock Campus, Attock, Pakistan, and associated with the Future Technology Research Center, National Yunlin University of Science and Technology, Douliou, Yunlin, Taiwan, for the research work. He has developed the Fractional least mean square algorithm and computational platform is formulated for the first time for solving fractional differential equation using artificial intelligence techniques during his Ph.D. studies. He has been author of more than 275 publications, out of which more than 225 are reputed journal publications with impact factor $ more than 850. He acts as a Resource Person and gives invited talks on many workshops and conferences held at the national level. His research interests include solving linear and nonlinear differential equation of arbitrary order, active noise control systems, fractional adaptive signal processing, nonlinear system identification, direction of arrival estimation, and bioinformatics problems.