Iranian Journal of Numerical Analysis and Optimization
Volume:13 Issue: 2, Spring 2023

In this manuscript, for approximation of solutions to equations that are nonlinear, a new class of two-point iterative structure that is based on a weight function involving two converging power series, is developed. For any method constructed from the developed class of methods, it requires three separate functions evaluation in a complete iteration cycle that is of order four convergence. Also, some well-known existing methods are typical members of the new class of methods. The numerical test on some concrete methods derived from the class of methods indicates that they are effective and competitive when employed in solving a nonlinear equation.

This article investigates the activity regimes of a realistic neuron model (as a slow-fast system). The authors study this model using the dynam-ical systems theory, for example, qualitative theory methods of slow-fast systems. The authors obtain the stability conditions of equilibria in leech heart interneurons under defined pharmacological conditions and following Hodgkin–Huxley formalism. Although in neuronal models, the membrane is usually considered  capacitance as a fixed parameter, the membrane ca-pacitance parameter is assumed as a control parameter to guarantee the existence of Hopf bifurcation using the Routh–Hurwitz criteria. The au-thors investigate the transition mechanism between the silent phase and tonic spiking mode. Furthermore, some simulations are provided using XPPAUT software for analytical results.

We apply the rational radial basis functions (RRBFs) method to solve the Allen–Cahn (A.C) equation, particularly when the equation has a so-lution with steep front or sharp gradients. We approximate the spatial derivatives by the RRBFs method. Then we apply an explicit, fourth-order Runge–Kutta method to advance the resulting semi-discrete system in time. It is well known that the A.C equation has a nonlinear stability feature, meaning that the free-energy functional is reduced by time. The presented method maintains the total energy reduction property of the A.C equation. In the end, five examples to confirm the efficiency and accuracyof the proposed method are provided.

In this paper,  a generalized version of the auto-convolution Volterra integral equation of the first kind as an ill-posed problem is studied. We apply the piecewise polynomial collocation method to reduce the numerical solution of this equation to a system of algebraic equations. According to the proposed numerical method, for $n=0$ and  $n=1,\ldots, N-1$, we obtain a  nonlinear and linear system, respectively. We have to distinguish between two cases, nonlinear and linear systems of algebraic equations. A double iteration process based on the modified Tikhonov regularization method is considered to solve the nonlinear algebraic equations. In this process, the outer iteration controls the evolution path of the unknown vector $U_0^{\delta}$ in the selected direction $\tilde{u}_0$, which is determined from the inner iteration process. For the linear case, we apply the Lavrentiev $\tilde{m}$ times iterated regularization method to deal with the ill-posed linear system. The validity and efficiency of the proposed method are demonstrated by several numerical experiments.

Despite the variety of methods available to solve nonlinear optimal con-trol problems, numerical methods are still evolving to solve these problems. This paper deals with the numerical solution of nonlinear optimal control affine problems by the interpolated variational iteration method, which was introduced in 2016 to improve the variational iteration method. For this purpose, the optimality conditions are first derived as a two-point bound-ary value problem and then converted to an initial value problem with the unknown initial values for costates. The speed and convergence of the method are compared with the existing methods in the form of three ex-amples, and the initial values of the costates are obtained by an efficient technique in each iteration.

This essay considers an optimal control problem (OCP) governed by a system of Fredholm integral equations (FIE). In this paper, collocation approach with utilizing Lagrange polynomials is introduced to transform the OCP into a nonlinear programming problem (NLP). An efficient op-timization method in Mathematica software is utilized to solve NLP. The convergence analysis is discussed, which show the theoretical structure behind the propounded technique under some assumptions. In this es-say, computational outcomes are given to demonstrate the adaptability, forthrightness, and relationship of the calculations manufactured. A prac-tical real-world problem involving hanging chain in classical mechanic is also dissolved utilizing the approach proposed.

In this article, we constructed a numerical scheme for singularly perturbed time-delay reaction-diffusion problems. For the discretization of the time derivative, we used the Crank-Nicolson method and a hybrid scheme, which is a combination of the fourth-order compact difference scheme and the cen-tral difference scheme on a special type of Shishkin mesh in the spatial di-rection. The proposed scheme is shown to be second-order accurate in time and fourth-order accurate with a logarithmic factor in space. The uniform convergence of the proposed scheme is discussed. Numerical investigations are carried out to demonstrate the efficacy and uniform convergence of the proposed scheme, and the obtained numerical results reveal the better per-formance of the present scheme, as compared with some existing methods in the literature.

The main goal of this work is to develop and analyze an accurate trun-cated stochastic Runge–Kutta (TSRK2) method to obtain strong numeri-cal solutions of nonlinear one-dimensional stochastic differential equations (SDEs) with continuous Hölder diffusion coefficients. We will establish the strong L1-convergence theory to the TSRK2 method under the local Lipschitz condition plus the one-sided Lipschitz condition for the drift co-efficient and the continuous Hölder condition for the diffusion coefficient at a time T and over a finite time interval [0, T ], respectively. We show that the new method can achieve the optimal convergence order at a finite time T compared to the classical Euler–Maruyama method. Finally, nu-merical examples are given to support the theoretical results and illustrate the validity of the method.

This paper proposes a numerical approach to solve a two-layer shal-low water formula with a slope and uneven bottom. The finite volume method (FVM) is applied to solve the shallow water model because the method is suitable for computational fluid dynamics problems. Rather than pointwise approximations at grid points, the FVM breaks the domain into grid cells and approximates the total integral over grid cells. The shallow water model is examined in two cases, the shallow water model in the steady state and the unsteady state. The quadratic upstream interpo-lation for convective kinetics (QUICK) is chosen to get the discretization of the space domain since it is a third-order scheme, which provides good accuracy, and this scheme proves its numerical stability. The advantage of the QUICK method is that the main coefficients are positive and satisfy the requirements for conservativeness, boundedness, and transportation. An explicit scheme is used to get the discretization of the time domain. Finally, the numerical solution of the steady state model shows that the flow remains unchanged. An unsteady-state numerical solution produces instability (wavy at the bottom layer). Moreover, the larger slope results in higher velocity and higher depth at the second layer.

The problem of the steady activation of a skeletal muscle is one of the ap-plicable phenomena in real life that can be modeled by a Volterra integral equation. The current research aims to investigate this problem by using an effective operational matrix-based method. For this purpose, the opera-tional matrix of integration is derived for the barycentric rational cardinal basis functions. Then, by utilizing the obtained operational matrix and without using any collocation points, the governing integral equation is re-duced to a system of nonlinear algebraic equations. Convergence analysis of the proposed numerical method is studied thoroughly. Moreover, the obtained numerical results based on the proposed method with acceptable computational times support the theoretical results and reveal the accuracy and efficiency of the method.