Abstract: The fractional differential equations involving different types of fractional derivatives are currently used in many fields of science and engineering. Therefore, the first purpose of this study is to investigate the qualitative properties including the stability, asymptotic stability, as well as Mittag–Leffler stability of solutions of fractional differential equations with the new generalized Hattaf fractional derivative, which encompasses the popular forms of fractional derivatives with non-singular kernels. These qualitative properties are obtained by constructing a suitable Lyapunov function. Furthermore, the second aim is to develop a new numerical method in order to approximate the solutions of such types of equations. The developed method recovers the classical Euler numerical scheme for ordinary differential equations. Finally, the obtained analytical and numerical results are applied to a biological nonlinear system arising from epidemiology.

Abstract: To minimize real-time errors in a Photovoltaic (PV) system performances must be forecasted through precise simulation design before continuing into a practical application. However, due to the scarcity of data in datasheets and the inherent transcendental connections are between PV current and PV voltage, to determining the Single Diode Model (SDM) parameters becomes a more challenging problems. This paper offers a simulated study of a SDM and Double Diode Model (DDM) solar PV system under various irradiation represents, and the performance was developed by incorporating an optimization-based Maximum Power Point (MPP) tracking techniques. According to the present simulation presented in this article, a mathematical model for a SDM/DDM as well as optimization methodologies has been estimated MATLAB platform. The present MPP circuit model designed and compared with BAT optimization algorithms. The nonlinear relationship between Voltage (V) - Current (I) and Voltage (V) –Power (W) acknowledged as characteristic curves for different temperature (∘c) and irradiance (W/m2) values are verified in numerical simulation results. MPP tracking power and efficiency are examined for maximum power (Pmax) to test the optimization based system. The simulation results show that the BAT optimization model was achieved the highest tracking efficiency better than other heuristic algorithms.

Abstract: This document is meant to be a practical introduction to the analytical and numerical manipulation of fermionic Gaussian systems. Starting from the basics, we move to relevant modern results and techniques, presenting numerical examples and studying relevant Hamiltonians, such as the transverse field Ising Hamiltonian, in detail. We finish introducing novel algorithms connecting fermionic Guassian states with matrix product states techniques. All the numerical examples make use of the free Julia package F_utilities.

Abstract: In this research article, we develop a powerful algorithm for numerical solutions to variable-order partial differential equations (PDEs). For the said method, we utilize properties of shifted Legendre polynomials to establish some operational matrices of variable-order differentiation and integration. With the help of the aforementioned operational matrices, we reduce the considered problem to a matrix type equation (equations). The resultant matrix equation is then solved by using computational software like Matlab to get the required numerical solution. Here it should be kept in mind that the proposed algorithm omits discretization and collocation which save much of time and memory. Further the numerical scheme based on operational matrices is one of the important procedure of spectral methods. The mentioned scheme is increasingly used for numerical analysis of various problems of differential as well as integral equations in previous many years. Pertinent examples are given to demonstrate the validity and efficiency of the method. Also some error analysis and comparison with traditional Haar wavelet collocations (HWCs) method is also provided to check the accuracy of the proposed scheme.

Abstract: In this work, we present some numerical results about variable order fractional differential equations (VOFDEs). For the said numerical analysis, we use Bernstein polynomials (BPs) with non-orthogonal basis. The method we use does not need discretization and neither collocation. Hence omitting the said two operations sufficient memory and time can be saved. We establish operational matrices for variable order integration and differentiation which convert the consider problem to some algebraic type matrix equations. The obtained matrix equations are then solved by Matlab 13 to get the required numerical solution for the considered problem. Pertinent examples are provided along with graphical illustration and error analysis to validate the results. Further some theoretical results for time complexity are also discussed.

Abstract: This paper presents a meshless finite point method (FPM) for the numerical analysis of the fractional cable equation. A second-order time discrete scheme is proposed to approximate both integer-order and fractional-order time derivatives. Then, based on the stabilized moving least squares approximation and the meshless smoothed gradients, a new implementation of the FPM is provided to enhance the accuracy and convergence rate in space. Theoretical error of the FPM is analyzed. Numerical results verify the efficiency of the method and show that the method can gain second-order accuracy in time and fourth-order accuracy in space.

Abstract: Abstract A second-order accurate (in time) and linear numerical scheme is proposed and analyzed for the nonlocal Cahn–Hilliard equation. The backward differentiation formula is used as the temporal discretization, while an explicit extrapolation is applied to the nonlinear term and the concave expansive term. In addition, an $O (\varDelta {t}^2)$ artificial regularization term, in the form of $A \varDelta _N (\phi ^{n+1} - 2 \phi ^n + \phi ^{n-1})$, is added for the sake of numerical stability. The resulting constant-coefficient linear scheme brings great numerical convenience; however, its theoretical analysis turns out to be very challenging, due to the lack of higher-order diffusion in the nonlocal model. In fact, a rough energy stability analysis can be derived, where an assumption on the $\ell ^\infty $ bound of the numerical solution is required. To recover such an $\ell ^\infty $ bound, an optimal rate convergence analysis has to be conducted, which combines a high-order consistency analysis for the numerical system and the stability estimate for the error function. We adopt a novel test function for the error equation, so that a higher-order temporal truncation error is derived to match the accuracy for discretizing the temporal derivative. Under the view that the numerical solution is actually a small perturbation of the exact solution, a uniform $\ell ^\infty $ bound of the numerical solution can be obtained, by resorting to the error estimate under a moderate constraint of the time step size. Therefore, the result of the energy stability is restated with a new assumption on the stabilization parameter $A$. Some numerical experiments are carried out to display the behavior of the proposed second-order scheme, including the convergence tests and long-time coarsening dynamics.

Abstract: Solving analytically intractable partial differential equations (PDEs) that involve at least one variable defined on an unbounded domain arises in numerous physical applications. Accurately solving unbounded domain PDEs requires efficient numerical methods that can resolve the dependence of the PDE on the unbounded variable over at least several orders of magnitude. We propose a solution to such problems by combining two classes of numerical methods: (i) adaptive spectral methods and (ii) physics-informed neural networks (PINNs). The numerical approach that we develop takes advantage of the ability of PINNs to easily implement high-order numerical schemes to efficiently solve PDEs and extrapolate numerical solutions at any point in space and time. We then show how recently introduced adaptive techniques for spectral methods can be integrated into PINN-based PDE solvers to obtain numerical solutions of unbounded domain problems that cannot be efficiently approximated by standard PINNs. Through a number of examples, we demonstrate the advantages of the proposed spectrally adapted PINNs in solving PDEs and estimating model parameters from noisy observations in unbounded domains.

Abstract: Abstract The primary goal of this article is to propose a new fractional boundary element technique for solving nonlinear three-temperature (3 T) thermoelectric problems. Analytical solution of the current problem is extremely difficult to obtain. To overcome this difficulty, a new numerical technique must be developed to solve such problem. As a result, we propose a novel fractional boundary element method (BEM) to solve the governing equations of our considered problem. Because of the advantages of the BEM solution, such as the ability to treat problems with complicated geometries that were difficult to solve using previous numerical methods, and the fact that the internal domain does not need to be discretized. As a result, the BEM can be used in a wide variety of thermoelectric applications. The numerical results show the effects of the magnetic field and the graded parameter on thermal stresses. The numerical results also validate the validity and accuracy of the proposed technique.

Abstract: The article presents experimental and numerical studies of bird models during impacts with rigid and deformable targets. The main aim of the studies is the validation of bird models in order to prepare them for the numerical simulation of bird impact against aircraft windshields and other parts of aircraft, thus improving the air transportation safety by providing cost-effective solutions for designing bird strike-resistant aircraft. The experimental investigations were conducted with a special set-up of a gas gun equipped with high-speed cameras, tensiometers and force sensors. The simulations were developed on the basis of LS-DYNA software by means of the SPH method for the bird model shape of the cylinder with hemispherical endings at the speed of 116 m/s. The results of studies into such things as the impact force, pressure and bird model deformation were compared. Moreover, the authors’ and other researchers’ results were assessed. It can be noted that the curves of the impact force obtained as a result of the numerical analysis correlated well with the experimental ones.

Abstract: Understanding, quantifying, and forecasting water flow and its behavior in environment is made possible by the use of computational hydraulics in con-junction with numerical models, which is one of the most powerful tools currently available. It is made up of simple to complex mathematical equations having linear and/or nonlinear elements, as well as ordinary and partial differential equations, and it is used to solve problems in many areas. In the vast majority of cases, it is not useful to reach analytical solutions to these mathematical equations using conventional methods. In these settings, mathematical models are solved by employing a variety of numerical algorithms and associated schemes. As a result, in this manuscript, we will cover the most fundamental numerical approach, the Finite Difference Method (FDM), in order to reformulate the governing equations for water and sediment flow from a system of partial differential equations to a system of linear equations. As part of our analysis into the inner workings of a computer program known as MIKE 21C, we will attempt to gain a better understanding of the hydrodynamic processes that take place in major rivers in Bangladesh. In addition to that, we will go over some of the most commonly used morphological studies that have been conducted on Bangladesh’s major rivers, including morphological solutions that have been developed in response to water supply con-cerns.

Abstract: Deep learning (DL) has become an integral part of solutions to various important problems, which is why ensuring the quality of DL systems is essential. One of the challenges of achieving reliability and robustness of DL software is to ensure that algorithm implementations are numerically stable. DL algorithms require a large amount and a wide variety of numerical computations. A naive implementation of numerical computation can lead to errors that may result in incorrect or inaccurate learning and results. A numerical algorithm or a mathematical formula can have several implementations that are mathematically equivalent, but have different numerical stability properties. Designing numerically stable algorithm implementations is challenging, because it requires an interdisciplinary knowledge of software engineering, DL, and numerical analysis. In this paper, we study two mature DL libraries PyTorch and Tensorflow with the goal of identifying unstable numerical methods and their solutions. Specifically, we investigate which DL algorithms are numerically unstable and conduct an in-depth analysis of the root cause, manifestation, and patches to numerical instabilities. Based on these findings, we launch DeepStability, the first database of numerical stability issues and solutions in DL. Our findings and DeepStability provide future references to developers and tool builders to prevent, detect, localize and fix numerically unstable algorithm implementations. To demonstrate that, using DeepStability we have located numerical stability issues in Tensorflow, and submitted a fix which has been accepted and merged in.