Neural network approaches based on new NCP-functions for solving tensor complementarity problem

Two new NCP-functions are constructed firstly in this paper. The main purpose is to accelerate the process of solution-finding for tensor complementarity problem, which is implemented by neural network methods based on the promising NCP-functions. Moreover, the stability properties of the proposed n...

Full description

Saved in:
Bibliographic Details
Published in:Journal of applied mathematics & computing 2021, Vol.67 (1-2), p.833-853
Main Authors: Xie, Ya-Jun, Ke, Yi-Fen
Format: Article
Language:English
Subjects:
Applied mathematics
Computational Mathematics and Numerical Analysis
Mathematical analysis
Mathematical and Computational Engineering
Mathematics
Mathematics and Statistics
Mathematics of Computing
Neural networks
Original Research
Stability
Tensors
Theory of Computation
Citations: Items that this one cites
Items that cite this one
Online Access:Get full text
Tags: Add Tag
No Tags, Be the first to tag this record!
Description
Summary:Two new NCP-functions are constructed firstly in this paper. The main purpose is to accelerate the process of solution-finding for tensor complementarity problem, which is implemented by neural network methods based on the promising NCP-functions. Moreover, the stability properties of the proposed neural networks are achieved via some theoretics and properties of generalization for linear and nonlinear complementarity problems. Plentiful numerical simulations demonstrate that the presented neural networks possess significantly better stability and comparable convergence rates than neural networks based on some existing NCP-functions.
ISSN:1598-5865
1865-2085
DOI:10.1007/s12190-021-01509-w