Vol. 4, No. 4 (2015): Regular Issue

The special issue of DNC “Fractional Dynamics and Systems with Power- Law Memory” contains papers related to presentations given at the 5th Conference on Nonlinear Science and Complexity, NSC’14, held on August 4-9, 2014 in Xi’an Jiaotong University, Xi’an, P. R. of China. Within this conference the authors proposed two mini-symposia: “Nonlinear Fractional Dynamics and Systems with Memory” and “Fractional Calculus Applications”.

In the last three decades Fractional Calculus (FC) became an area of intense research and development. The accompanying poster illustrates the present day major achievements in the application of FC in physics, engineering and biology.

In this paper we extend the notion of an α-family of maps to discrete systems defined by simple difference equations with the fractional Caputo difference operator. The equations considered are equivalent to maps with falling factorial-law memory which is asymptotically power-law memory. We introduce the fractional difference Universal, Standard, and Logistic α- Families of Maps and propose to use them to study general properties of discrete nonlinear systems with asymptotically power-law memory.

Terrorist attacks are catastrophic events often accompanied by a large number of human losses. The statistics of these casualties can be approximated by Power Law (PL) distributions. In this paper we analyze a dataset of terrorist events by means of PL distributions and Pseudo Phase Plane (PPP) technique. We consider worldwide events grouped into 13 geographical regions. First, for each region, we approximate the empirical data by PL functions and we analyze the emerging PL parameters. Second, we model the dataset as time-series and interpret the data as the output of a dynamical system. For each region, we compute the correlation coefficient to find the optimal time delay for reconstructing the PPP. Third, we compare the PPP curves using clustering tools in order to unveil relationships among the data.

This paper deals with a previously ignored problem that how to find the memory (initialization function) of fractional order system by using the recent sampled input-output data. A novel and practical strategy is proposed to estimate the initialization function, which adapts to all system parameters but fractional order. To implement this method, a P-type order learning approach is introduced to identify the system order separably and accurately, thanks to the fractional order sensitivity function. The initialization response is computed through an iterative learning identification strategy that guarantees the accuracy and adaptiveness simultaneously. Along with the estimations of order and initialization response, a practical piecewise identification criterion of initialization function is established by using the least squares and instrumental variable methods. The above strategy is available for both Caputo and Riemann-Liouville fractional order systems, where the initial values are applied rather than the initial conditions. Two illustrated examples are provided to support the conclusions.

The optimal control problem for linear dynamic systems of fractional order with lumped and distributed parameters is investigated. This problem is reduced to the classical moment problem. This paper validates the conditions making possible to formulate and resolve the obtained moment problem. Some particular cases of fractional-order systems are discussed. The explicit solutions for the problems of optimal control were obtained in case of systems with lumped parameters. In case of system with distributed parameters an approximate solution analyzed for moment problem. In particular, this paper studies the problem to minimize the norm of control for the assigned time interval and the problem of control with the minimal time of the object transition into the desirable state with the given limitation of the norm of control.

This paper proposes sliding mode control for the 4-D fractional order Lorenz-Stenflo hyperchaotic system. Two methods are utilized: one is based on the frequency distributed model of fractional integral operator; and the other is based on the Mittag-Leffler stability theorem and the Caputo operator property. Both of the two methods involve two steps: firstly, constructing a fractional order sliding surface; secondly, designing a single sliding control law for suppression of the nominal plant. Numerical simulations are carried out to verify the efficiency of the theoretical results.

In this paper, a fractional-order drive-response complex network model with different order nodes is proposed for the first time. To achieve the hybrid projective synchronization (HPS) of drive-response complex network with different orders, a general strategy is proposed and effective controllers for hybrid projective synchronization are designed. The fractional operators are introduced into the controller to transform problem into synchronization problem between drive-response complex network with identical orders. Numerical simulation results which are carried show that the method is easy to implement and reliable for synchronizing the driveresponse fractional-order complex networks.

In this paper, we investigate the existence of solutions for a four-point nonlocal boundary value problem of nonlinear impulsive differential equations of fractional order α ∈ (2,3]. By using some well known fixed point theorems, sufficient conditions for the existence of solutions are established. Some illustrative examples are also discussed.

First we discuss some crucial factors that determine numerical calculations accuracy of the Gr¨unwald-Letnikov formula and its equivalent Horner form. Then we introduce simplified variants of both formulas and the concept of the calculation tail. We analyze the utility of its length for mitigation of a time and a memory shortages influence on the accuracy in realtime microprocessor calculations. Credibility of the conclusions is lent by the comparison of the results obtained on a PC and on a real-time DSP system.

We propose to replace the Gauss Quadratures with a numerical integration method known as the Double Exponential (DE) Formula. The numerical quadrature built upon it is at least equivalently accurate and much simpler to customize and apply in situations when tabulated values of the Gauss Quadratures’ nodes and weights can not be applied. The DE Formula was developed for integrals with endpoint singularities. However, we confirm that it can be successfully applied to any integral and interval, for which the Gauss Quadratures have been usually selected. To remain compact, the following presentation focuses only on the most difficult integrals, e.g. the improper integrals and the integrals with endpoint singularities. The main part of the paper consists of the calculations accuracy comparison between numerical quadrature based upon the DE Formula and the Gauss-Laguerre, the Gauss-Hermite or Gauss-Chebyshev Quadratures.