DocumentCode :
572484
Title :
Zhang fractals yielded via solving nonlinear equations by discrete-time complex-valued ZD
Author :
Wu, Huarong ; Li, Fen ; Li, Zhan ; Zhang, Yunong
Author_Institution :
Sch. of Inf. Sci. & Technol., Sun Yat-sen Univ., Guangzhou, China
fYear :
2012
fDate :
15-17 Aug. 2012
Firstpage :
1
Lastpage :
6
Abstract :
In this paper, a novel kind of new fractals, named Zhang fractals, is yielded by using the discrete-time complex-valued Zhang dynamics (DTCVZD) to solve the nonlinear equations in the complex domain. Such a novel DTCVZD model is designed based on the elimination of an indefinite complex-valued error function, instead of a square-based non-negative energy function associated with the discrete-time complex-valued gradient-based dynamics (DTCVGD). Comparing with the well-known (generalized) Newton fractals (i.e., the famous fractals generated by the well-known Newton iteration), we find that the novel Zhang fractals synthesized by the proposed DTCVZD model incorporate such Newton fractals as special cases. The Zhang fractals generated by the novel DTCVZD model are completely different from Newton fractals. The DTCVZD model using different types of activation functions can be seen as a new iterative algorithm to generate new fractals, i.e., Zhang fractals.
Keywords :
fractals; gradient methods; nonlinear equations; DTCVGD; DTCVZD model; Newton fractals; Zhang fractals; activation functions; discrete-time complex-valued Zhang dynamics; discrete-time complex-valued gradient-based dynamics; indefinite complex-valued error function; iterative algorithm; nonlinear equations; square-based nonnegative energy function; Computational modeling; Educational institutions; Fractals; Mathematical model; Neural networks; Nonlinear equations; Fractals; Zhang dynamics (ZD); complex-valued; discrete-time; nonlinear equations;
fLanguage :
English
Publisher :
ieee
Conference_Titel :
Automation and Logistics (ICAL), 2012 IEEE International Conference on
Conference_Location :
Zhengzhou
ISSN :
2161-8151
Print_ISBN :
978-1-4673-0362-0
Electronic_ISBN :
2161-8151
Type :
conf
DOI :
10.1109/ICAL.2012.6308160
Filename :
6308160
Link To Document :
بازگشت