A Nonlinear Dimensional Analysis Framework Using Fuzzy Graph Theory and Deep Neural Networks for Quantum-Intelligent Systems

Authors

  • Dr. Shanmugasundaram M Author
  • B. Balavinothini Author
  • P. Sudheer Author
  • Dr.K. Sathishkumar Author
  • Dr. K. Radha Rani Author
  • Elangovan Muniyandy Author
  • Dr.P.Dharmendra Kumar Author
  • R. Naveenkumar Author

Keywords:

Nonlinear dimensional analysis; fuzzy graph theory; deep neural networks; quantum-inspired intelligence; minimax optimization; nonlinear systems.

Abstract

Nonlinear systems with high dimensions are highly challenging to model due to complex inter-dimensional interactions, imprecise measurements and time-varying interactions among system variables. Traditional non linear learning techniques are usually applied in dimensions separately or on given representations, therefore are limited in uncertainty and varying relations structures. This paper provides the hypothesis of a nonlinear dimensional analysis framework to model the intelligent system robustly in quantum-inspired combining fuzzy graph theory with deep neural networks. The uncertain system states are described by using fuzzy representations and the weighted graph structure is utilized to represent nonlinear relationships between interacting dimensions. The dimension-aware deep graph learning mechanism learns adaptive weights of dimensions and obtains smaller nonlinear representations. Then, the learned features are converted by means of a pattern representation inspired by quantum computing for the purpose of distinguishing between stable, transitional and complex state of systems. The learning process is proposed to be a minimax optimization problem to increase the bounded perturbations' resistance whose loss is minimized jointly for the prediction and classification. Computational experiments with varying levels of uncertainties and dimensions are conducted to compare the proposed framework with the traditional neural as well as graph models. Root mean square error (RMSE), mean absolute error (MAE), coefficient of determination ((R^2)), classification accuracy and Macro F1-score are used to evaluate performance. Results have proved that the above framework can be applied accurately and uncertainty in intelligent high dimensional systems modeling and strong pattern identification with nonlinearity.

Downloads

Download data is not yet available.

Downloads

Published

2026-09-22

How to Cite

A Nonlinear Dimensional Analysis Framework Using Fuzzy Graph Theory and Deep Neural Networks for Quantum-Intelligent Systems. (2026). Minimax Theory and Its Applications, 11(2s), 841-848. https://journalmta.com/index.php/jmta/article/view/457