Current Issue

2020, Vol. 40 No. 3   Published date: 30 September 2020
Previous Issue    Next Issue

• Dynamic Filtering Algorithms Based on State Constrained with Uncertainty

  2020, 40(3): 1-10. doi:

  Abstract ( 1293 )   PDF (1166KB) ( 385 )   　　

   Dynamic Filtering models were based on state constrained with uncertainty by regarding uncertainty as state constraint, incorporated into the adjustment models. The unconstrained filtering algorithm is proposed according to unconstrained adjustment model, which is equal to standard kalman filtering algorithm. The inequality state constrained filtering algorithm and ellipsoid state constrained filtering algorithm are provided as the solution of state constrained adjustment models. With simulation calculation, it is compared that the results of different dynamic filtering algorithms based on state constrained with uncertainty are compared. The results show that dynamic filtering algorithms based on state constrained with uncertainty are better than kalman filtering algorithm, obtained simplicity and effectivity. 

• State-feedback Stabilization of a Class of Non-autonomous Stochastic Time-delay Systems

  2020, 40(3): 11-21. doi:

  Abstract ( 1206 )   PDF (837KB) ( 275 )   PDF（mobile） (837KB) ( 0 )   　　

  The paper studies the state feedback stabilization of a class of non-autonomous stochastic time-delay sys- tems. On the one hand, due to the intrinsic characteristics of system model (e.g., complex nonlinearity, randomness and non-autonomous), a time-varying feedback controller together with a state feedback control scheme is designed via the backstepping method. On the other hand, based on Razumikhin-type stochastic stability theorem, we analyze the global asymptotic stability in the sense of probability for the closed-loop system with weaker conditions on time-delay.  Final- ly, two simulation examples are presented to illustrate the effectiveness of the proposed feedback controller and control scheme.

• Augmented truncation approximations for countable Markov chains

  2020, 40(3): 22-39. doi:

  Abstract ( 1188 )   PDF (762KB) ( 366 )   　　

• Numerical solution for a semilinear parabolic blowup-combustion model

  2020, 40(3): 40-53. doi:

  Abstract ( 1309 )   PDF (1014KB) ( 289 )   　　

  In this paper, we study an efficient finite difference method based on constructed mesh for a class of semilinear parabolic equation arising from combustion process. Due to the influence of exponential source term, the considered model may theoretically describe the rigid ignition process in combustion. To catch the blowup and almost complete blowup phenomena, we introduce a cut-off level constant in the indicator attached to source term, so that the blowup position and moment can be approximated with less effort. Finally, numerical simulations with various settings of parameters and initial conditions are given to illustrate the excellent performance of numerical scheme, as well as fruitful dynamics of the blowup-combustion model.

• An Improved Straight Walk Algorithm for Point Location in a Tetrahedral Mesh

  2020, 40(3): 54-64. doi:

  Abstract ( 1244 )   PDF (909KB) ( 295 )   　　

  The straight walk algorithm is commonly-used in searching for a mesh cell containing a query point in a given large scale unstructured mesh of a bounded domain. It could be used in the computational geometry and several other fields, including finite element solutions of partial differential equations. However, this algorithm does not work in some degenerate cases, such as the intersection point coincides with the vertex. In this paper, as an improvement, a new straight walk algorithm is presented to make it work for degenerate (or singular) cases in a tetrahedral mesh, which is well verified in our numerical tests.

• Drone Object Detection Based on Retina Magnocelluar Pathway Model and Deep Learning

  2020, 40(3): 65-76. doi:

  Abstract ( 1387 )   PDF (1026KB) ( 587 )   　　

• On the Extreme of the Degree Deviation of the Maximal Outerplanar Graphs

  2020, 40(3): 77-84. doi:

  Abstract ( 1289 )   PDF (786KB) ( 289 )   　　

• Research on Consumption and Optimal Investment Strategies  Based on Hyperbolic Discount Method

  2020, 40(3): 85-93. doi:

  Abstract ( 1304 )   PDF (725KB) ( 413 )   　　

   In this article, we mainly discuss the optimal amount of investment for risky assets in stochastic hyperbolic discounting, assuming that investors’ consumption behavior is a Brownian motion. Based on the Hamilton-Jacobi-Bellman equation, the optimal investment portfolio with the constant absolute risk aversion investor is calculated, and the approximate solution of equation is given. Moreover, we analyzed some important properties of risky asset investment when consumption obeys the Wiener process, and then studied the relationship between consumption behavior and risky asset investment behavior.

• A Dependent Risk Model with Refunding and Dividend Strategy under Sparse Processes

  2020, 40(3): 94-100. doi:

  Abstract ( 1154 )   PDF (657KB) ( 229 )   　　

• Forecast of High Transfer of Listed Companies Based on Portfolio Model#br#

  2020, 40(3): 101-109. doi:

  Abstract ( 1260 )   PDF (981KB) ( 352 )   　　

  Based on the data of 3465 listed companies in the Chinese market in 7 years,this paper firstly extracted 43 factors by using random forest algorithm,and then used Lasso method to select the characteristics of the 43 factors selected,and finally selected 11 important factors.Then logistic regression is used to build the first prediction model,and then the decision tree model is used to build the second prediction model. Finally,the combination model based on the loss function to determine the weight is linear combination of the first prediction model and the second prediction model to build the combination model. The empirical results show that the prediction accuracy of the combined model is 1.39% higher than that of the single model.

• A Decision-making Method of the Airport Group Informatization Project Implementation Sequence Based on the Synergy Effect

  2020, 40(3): 110-120. doi:

  Abstract ( 1245 )   PDF (863KB) ( 225 )   　　

• Fast Modeling of Ultrasonic Transducer Sound Field and Optimization Design of Acoustic Hologram

  2020, 40(3): 121-128. doi:

  Abstract ( 1208 )   PDF (855KB) ( 894 )   　　

  In this paper, in order to improve the speed of finite element simulation software such as COMSOL to simulate the three-dimensional (3D) sound field of ultrasonic transducer (UT), the fast modelling method for the sound field of UT is developed based on the angular spectrum method (ASM). In addition, the control method for the sound field of UT is proposed by combining the ASM with Gerchberg-Saxton iterative algorithm. Based on the calculated phase difference, the acoustic hologram of UT can be fabricated by using the 3D printing technology to accurately control the sound field of UT.