Secrecy Capacity Maximization in Wireless Wiretap Channel: A Neurodynamic Optimization Approach

Abstract

This paper addresses the secure transmission problem of privacy information over a fading channel with an eavesdropper. A neural network model is proposed for solving the secrecy capacity maximization problems in real time. Unlike conventional power allocation strategies, a neurodynamic secure transmission approach is provided by the relation between KKT (Karush-Kuhn-Tucker) optimality conditions and the equilibrium point of a neural network. The transient behaviour of neural networks are showed, and the effectiveness of the neurodynamic approach is substantiated with a secrecy capacity maximization problem.