A [formula omitted]-power neurodynamic approach to distributed nonconvex optimization.

In this paper, a neurodynamic optimization approach based on a p -power transformation Lagrangian function is developed for distributed nonconvex optimization. A new Lagrangian function is proposed to eliminate dual gaps of nonconvex problems, and a distributed average tracking approach is developed for estimating global objective function value. Based on the Lagrangian function and the distributed average tracking approach, a neurodynamic model is developed for distributed nonconvex optimization, and its convergence to a local minimum is proven. Two numerical examples are provided to demonstrate the validity and effectiveness of the proposed approach. • A Lagrangian function with a p -power transformation is designed. • A neurodynamic model for distributed nonconvex optimization is developed. • The convergence of the neurodynamic model to a local minimum is proven. • Two numerical examples are provided to demonstrate the validity of the approach. [ABSTRACT FROM AUTHOR]