Abstract
Recently, zeroing dynamics (ZD) method has shown wonderful effect in solving time-varying problems. Generally, the continuous-time ZD model should be discretized as a discrete-time algorithm for numerical computation. A good choice is to apply Zhang et al. discretization (ZeaD) formulas, which are effective time-discretization formulas. Actually, ZeaD formulas can also be applied to obtaining various discrete-time algorithms, not only the ZD algorithms. In the previous work, some piecemeal ZeaD formulas have been proposed and investigated. However, the relationship between the time-instant number and precision of ZeaD formulas is not found, which is emphatically investigated in this paper. Specifically, the ZeaD formulas from two to nine instants are investigated, and the general ZeaD formula groups are studied. Two-instant and three-instant ZeaD formula groups have linear precision at most. Four-instant ZeaD formula group has quadratic precision at most. Five-instant and six-instant ZeaD formula groups have cubic precision at most. Seven-instant and eight-instant ZeaD formula groups have quartic precision at most. Nine-instant ZeaD formula group has quintic precision at most. Theoretical analyses are presented to substantiate the relationship. Moreover, the ZeaD formulas as well as ZD method are applied to solving time-varying quadratic optimization problem, and the numerical results verify the effectiveness of ZeaD formulas.
Similar content being viewed by others
References
Na, J., Xing, Y., Costa-Castello, R.: Adaptive estimation of time-varying parameters with application to roto-magnet plant. IEEE Trans. Syst., Man, Cybern. -Syst. (to be published. https://doi.org/10.1109/TSMC.2018.2882844https://doi.org/10.1109/TSMC.2018.2882844)
Liu, Y.J., Gong, M., Liu, L., Tong, S., Chen, C.L.P.: Fuzzy observer constraint based on adaptive control for uncertain nonlinear MIMO systems with time-varying state constraints. IEEE Trans. Cybern. (to be published. https://doi.org/10.1109/TCYB.2019.2933700)
Li, W.: A recurrent neural network with explicitly definable convergence time for solving time-variant linear matrix equations. IEEE Trans. Ind. Inform. 14(12), 5289–5298 (2018)
Stanimirovic, P.S., Katsikis, V.N., Li, S.: Higher-order ZNN dynamics. Neural Process Lett. 51(12), 697–721 (2020)
Stanimirovic, P.S., Katsikis, V.N., Li, S.: Integration enhanced and noise tolerant ZNN for computing various expressions involving outer inverses. Neurocomputing 329, 129–143 (2019)
Tan, Z., Xiao, L., Chen, S., Lv, X.: Noise-tolerant and finite-time convergent ZNN models for dynamic matrix Moore-Penrose inversion. IEEE Trans. Ind. Inform. 16(3), 1591–1601 (2020)
Guo, D., Yan, L., Nie, Z.: Design, analysis, and representation of novel five-step DTZD algorithm for time-varying nonlinear optimization. IEEE Trans. Neural Netw. Learn. Syst. 29(9), 4248–4260 (2018)
Zhang, Z., Lu, Y., Zheng, L., Li, S., Li, Y.: A new varying-parameter convergent-differential neural-network for solving time-varying convex QP problem constrained by linear-equality. IEEE Trans. Autom. Control 63(12), 4110–4125 (2018)
Zhang, Y., Yi, C.: Zhang Neural Networks and Neural-Dynamic Method. Nova Science Publishers Inc, New York (2011)
Zhang, Y., Xiao, L., Ruan, G., Li, Z.: Continuous and discrete time Zhang dynamics for time-varying 4th root finding. Numer. Algorithms 57(1), 35–51 (2011)
Guo, D., Nie, Z., Yan, L.: Theoretical analysis, numerical verification and geometrical representation of new three-step DTZD algorithm for time-varying nonlinear equations solving. Neurocomputing 214, 516–526 (2016)
Guo, D., Lin, X., Su, Z., Sun, S., Huang, Z.: Design and analysis of two discrete-time ZD algorithms for time-varying nonlinear minimization. Numer. Algorithms 77, 23–36 (2018)
Qiu, B., Zhang, Y., Guo, J., Yang, Z., Li, X.: New five-step DTZD algorithm for future nonlinear minimization with quartic steady-state error pattern. Numer. Algorithms 81, 1043–1065 (2019)
Chen, D., Li, S., Lin, F.J., Wu, Q.: New super-twisting zeroing neural-dynamics model for tracking control of parallel robots: A finite-time and robust solution. IEEE Trans. Cybern. 50(6), 2651–2660 (2020)
Li, S., Guo, Y.: Discrete-time consensus filters for average tracking of time-varying inputs on directed switching graphs. Asian J. Control 20 (2), 919–934 (2018)
Zhang, Y., Li, S., Liao, L.: Consensus of high-order discrete-time multiagent systems with switching topology. IEEE Trans. Syst. Man Cybern. -Syst. (to be published. https://doi.org/10.1109/TSMC.2018.2882558https://doi.org/10.1109/TSMC.2018.2882558)
Chen, J., Zhang, Y.: Continuous and discrete zeroing neural dynamics handling future unknown-transpose matrix inequality as well as scalar inequality of linear class. Numer. Algorithms 83, 529–547 (2020)
Liu, Y.J., Li, S., Tong, S., Chen, C.L.P.: Adaptive reinforcement learning control based on neural approximation for nonlinear discrete-time systems with unknown nonaffine dead-zone input. IEEE Trans. Neural Netw. Learn. Syst. 30(1), 295–305 (2019)
Mathews, J.H., Fink, K.D.: Numerical Methods Using Matlab. Prentice-Hall, Englewood Cliffs (2004)
Zhang, Y., Li, Z., Guo, D., Ke, Z., Chen, P.: Discrete-time ZD, GD and NI for solving nonlinear time-varying equations. Numer. Algorithms 64 (4), 721–740 (2013)
Jin, L., Zhang, Y.: Continuous and discrete Zhang dynamics for real-time varying nonlinear optimization. Numer. Algorithms 73(1), 115–140 (2016)
Li, J., Mao, M., Uhlig, F., Zhang, Y.: A 5-instant finite difference formula to find discrete time-varying generalized matrix inverses, matrix inverses, and scalar reciprocals. Numer. Algorithms 81, 609–629 (2019)
Zhang, Y., Liu, X., Ling, Y., Yang, M., Huang, H.: Continuous and discrete zeroing dynamics models using JMP function array and design formula for solving time-varying Sylvester-transpose matrix inequality. Numer. Algorithms (to be published. https://doi.org/10.1007/s11075-020-00946-1)
Zhang, Y., Zhu, M., Hu, C., Li, J., Yang, M.: Euler-precision general-form of Zhang others discretization (ZeaD) formulas, derivation, and numerical experiments. In: Proc. of Chinese Control and Decision Conference (CCDC), Shenyang, China, pp. 6262–6267 (2018)
Hu, C., Kang, X., Zhang, Y.: Three-step general discrete-time Zhang neural network design and application to time-variant matrix inversion. Neurocomputing 306, 108–118 (2018)
Li, J., Zhang, Y., Mao, M.: General square-pattern discretization formulas via second-order derivative elimination for zeroing neural network illustrated by future optimization. IEEE Trans. Neural Netw. Learn. Syst. 30(3), 891–901 (2019)
Zhang, Y., He, L., Hu, C., Guo, J., Li, J., Shi, Y.: General four-step discrete-time zeroing and derivative dynamics applied to time-varying nonlinear optimization. J. Comput. Appl. Math. 347, 314–329 (2019)
Hu, C., Zhang, Y., Kang, X.: General and improved five-step discrete-time zeroing neural dynamics solving linear time-varying matrix equation with unknown transpose. Neural Process Lett. 51, 1715–1730 (2020)
Zhang, Y., Guo, J., He, L., Shi, Y., Hu, C.: Any ZeaD formula of six instants having no quartic or higher precision with proof. In: Proc. of International Conference on Systems and Informatics (ICSAI), Nanjing, China, pp. 681–685 (2018)
Sun, M., Wang, Y.: General five-step discrete-time Zhang neural network for time-varying nonlinear optimization. Bull. Malays. Math. Sci. Soc. 43, 1741–1760 (2020)
Yang, M., Zhang, Y., Hu, H., Qiu, B.: General 7-instant DCZNN model solving future different-level system of nonlinear inequality and linear equation. IEEE Trans. Neural Netw. Learn. Syst. 31(9), 3204–3214 (2020)
Suli, E., Mayers, D.F.: An Introduction to Numerical Analysis. Cambridge University Press, Cambridge (2003)
Oppenheim, A.V.: Discrete-time Signal Processing, 3rd edn. Pearson Higher Education, New Jersey (2010)
Ogata, K.: Modern Control Engineering, 4th edn. Prentice Hall, New Jersey (2001)
Qin, S., Xue, X.: A two-layer recurrent neural network for nonsmooth convex optimization problems. IEEE Trans. Neural Netw. Learn. Syst. 26(6), 1149–1160 (2015)
Qin, S., Le, X., Wang, J.: A neurodynamic optimization approach to bilevel linear programming. IEEE Trans. Neural Netw. Learn. Syst. 28(11), 2580–2591 (2017)
Liu, N., Qin, S.: A neurodynamic approach to nonlinear optimization problems with affine equality and convex inequality constraints. Neural Netw. 109, 147–158 (2019)
Liu, N., Qin, S.: A novel neurodynamic approach to constrained complex-variable pseudoconvex optimization. IEEE Trans. Cybern. 49(11), 3946–3956 (2019)
Qin, S., Yang, X., Xue, X., Song, J.: A one-layer recurrent neural network for pseudoconvex optimization problems with equality and inequality constraints. IEEE Trans. Cybern. 47(10), 3063–3074 (2017)
Xiao, L.: A nonlinearly-activated neurodynamic model and its finite-time solution to equality-constrained quadratic optimization with nonstationary coefficients. Appl. Soft Comput. 40, 252–259 (2016)
Xiao, L., Li, K., Duan, M.: Computing time-varying quadratic optimization with finite-time convergence and noise tolerance: A unified framework for zeroing neural network. IEEE Trans. Neural Netw. Learn. Syst. 30(11), 3360–3369 (2019)
Xiao, L., Li, S., Yang, J., Zhang, Z.: A new recurrent neural network with noise-tolerance and finite-time convergence for dynamic quadratic minimization. Neurocomputing 285, 125–132 (2018)
Miao, P., Shen, Y., Huang, Y., Wang, Y.W.: Solving time-varying quadratic programs based on finite-time Zhang neural networks and their application to robot tracking. Neural Comput. Appl. 26(3), 693–703 (2015)
Zhang, Y., Gong, H., Yang, M., Li, J., Yang, X.: Stepsize range and optimal value for Taylor-Zhang discretization formula applied to zeroing neurodynamics illustrated via future equality-constrained quadratic programming. IEEE Trans. Neural Netw. Learn. Syst. 30(3), 959–966 (2019)
Zhang, Y., Qi, Z., Li, J., Qiu, B., Yang, M.: Stepsize domain confirmation and optimum of ZeaD formula for future optimization. Numer. Algorithms 81, 561–574 (2018)
Funding
The work is aided by the National Natural Science Foundation of China (with numbers 61976230 and 61673402), the Natural Science Foundation of Guangdong Province (with number 2017A030311029), the Project Supported by Guangdong Province Universities and Colleges Pearl River Scholar Funded Scheme (with number 2018), the Research Fund Program of Guangdong Key Laboratory of Modern Control Technology (with number 2017B030314165), and also the Key-Area Research and Development Program of Guangzhou (with number 202007030004).
Author information
Authors and Affiliations
Corresponding author
Additional information
Publisher’s note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
Appendices
Appendix: 1. Proof of three-instant ZeaD formula group
Evidently, when m = 3, three-instant ZeaD formula can be presented as
According to (2), if p = 2, one has
i.e.,
In this case, the characteristic equation is
whose roots are γ0 = 1 and γ1 = − 1. Hence, the synthesized formula is not enough 0-stable [32], i.e., three-instant ZeaD formula cannot have O(τ2) precision.
Meanwhile, if p = 1, one has
i.e.,
In this case, the characteristic equation is
By applying bilinear transformation [33, 34], i.e., γ = (ω + 1)/(ω − 1), one gets
From Routh stability criterion [33, 34], the 0-stable condition is a0 > − 1/2. Besides, a0≠ 0 should be satisfied. The proof is thus completed.
Appendix: 2. Proof of four-instant ZeaD formula group
When m = 4, four-instant ZeaD formula can be presented as
According to (2), if p = 3, one has
i.e.,
In this case, the characteristic equation is
whose roots are γ0 = 1, \(\gamma _{1}=(\sqrt {33}-5)/4\), and \(\gamma _{2}=(-\sqrt {33}-5)/4\). Evidently, |γ2| > 1. Hence, the synthesized formula is not 0-stable [32], i.e., four-instant ZeaD formula cannot have O(τ3) precision.
Meanwhile, if p = 2, one has
i.e.,
In this case, the characteristic equation is
By applying bilinear transformation [33, 34], i.e., γ = (ω + 1)/(ω − 1), one gets
From Routh stability criterion [33, 34], the 0-stable condition is a0 < 0. The general four-instant ZeaD formula (5) is obtained. The proof is thus completed.
Appendix: 3. Proof of five-instant ZeaD formula group
When m = 5, five-instant ZeaD formula can be presented as
According to (2), if p = 4, one has
i.e.,
In this case, the characteristic equation is
By applying bilinear transformation [33, 34], i.e., γ = (ω + 1)/(ω − 1), one gets the transformed characteristic equation
Because there is a negative coefficient − 8/3, the synthesized formula is not 0-stable [33, 34]. That is, five-instant ZeaD formula cannot have O(τ4) precision.
Meanwhile, if p = 3, one has
i.e.,
In this case, the characteristic equation is
By applying bilinear transformation [33, 34], i.e., γ = (ω + 1)/(ω − 1), one gets
From Routh stability criterion [33, 34], the 0-stable condition is 1/12 < a0 < 1/6. The general five-instant ZeaD formula (6) is obtained. The proof is thus completed.
Appendix: 4. Proof of six-instant ZeaD formula group
When m = 6, six-instant ZeaD formula can be presented as
According to (2), if p = 4, one has
i.e.,
In this case, the characteristic equation is
By applying bilinear transformation [33, 34], i.e., γ = (ω + 1)/(ω − 1), one gets transformed characteristic equation
Because there is a negative coefficient − 2, the synthesized formula is not 0-stable [33, 34]. That is, six-instant ZeaD formula cannot have O(τ4) precision.
Meanwhile, if p = 3, one has
i.e.,
In this case, the characteristic equation is
By applying bilinear transformation [33, 34], i.e., γ = (ω + 1)/(ω − 1), one gets
From Routh stability criterion [33, 34], the 0-stable condition is presented as the group form of
By analyzing and plotting, the condition can be reduced to
The general six-instant ZeaD formula (7) is obtained. The proof is thus completed.
Appendix: 5. Proof of seven-instant ZeaD formula group
When m = 7, seven-instant ZeaD formula can be presented as
According to (2), if p = 5, one has
i.e.,
In this case, the characteristic equation is
By applying bilinear transformation [33, 34], i.e., γ = (ω + 1)/(ω − 1), one gets transformed characteristic equation
Because there is a negative coefficient − 94/15, the synthesized formula is not 0-stable [33, 34]. That is, seven-instant ZeaD formula cannot have O(τ5) precision.
Meanwhile, if p = 4, one has
i.e.,
In this case, the characteristic equation is
By applying bilinear transformation [33, 34], i.e., γ = (ω + 1)/(ω − 1), one gets
From Routh stability criterion [33, 34], the 0-stable condition is in the group form as follows:
By analyzing and plotting, the condition can be reduced to
The general seven-instant ZeaD formula (8) is obtained. The proof is thus completed.
Rights and permissions
About this article
Cite this article
Yang, M., Zhang, Y. & Hu, H. Relationship between time-instant number and precision of ZeaD formulas with proofs. Numer Algor 88, 883–902 (2021). https://doi.org/10.1007/s11075-020-01061-x
Received:
Accepted:
Published:
Issue Date:
DOI: https://doi.org/10.1007/s11075-020-01061-x