Z-type and G-type ZISR (Zhang inverse square root) solving

A class of neural dynamics, called Zhang dynamics (ZD), has been proposed for online solution of various time-varying problems. In this paper, Z-type and G-type models, including continuous-time and discrete-time Z-type models, are proposed and simulated for solving the time-varying inverse square r...

Full description

Saved in:
Bibliographic Details
Main Authors: Yunong Zhang, Zhen Li, Yunjia Xie, Hongzhou Tan, Pei Chen
Format: Conference Proceeding
Language:English
Subjects:
Computational modeling
Convergence
Hardware
Heuristic algorithms
Integrated circuit modeling
Mathematical model
Problem-solving
Online Access:Request full text
Tags: Add Tag
No Tags, Be the first to tag this record!
Description
Summary:A class of neural dynamics, called Zhang dynamics (ZD), has been proposed for online solution of various time-varying problems. In this paper, Z-type and G-type models, including continuous-time and discrete-time Z-type models, are proposed and simulated for solving the time-varying inverse square root (or termed, Zhang inverse square root, ZISR) problem. Note that Z denotes Zhang and G denotes gradient. Moreover, the simplified Z-type models are generated for solving the static ISR (inverse square root) problem and the relationship between the Z-type models and Newton-Raphson iteration (NRI) is discovered. Through illustrative examples, the efficacy and superiority of the proposed Z-type models for time-varying and static ISR computation are verified.
DOI:10.1109/ICICIP.2013.6568053