A. M. Mukhamedov, “Application of simulated annealing algorithm to optimize parameters of differential neural network-based identifier of platform dynamics”, Intelligent systems. Theory and applications, 2021, Volume 25, Issue 4,Pages <nobr>289

Part 6. Intelligent control, robots and biomechatronics systems

Application of simulated annealing algorithm to optimize parameters of differential neural network-based identifier of platform dynamics

A. M. Mukhamedov^ab

^a Lomonosov Moscow State University
^b Center "Supersonic"

Abstract: Parameters of activation functions contribute a lot to the performance of differential neural network-based identifiers. An approach to optimize these parameters is proposed and demonstrated for the problem of identification of platform dynamics.

Keywords: differential neural networks, global optimization, simulated annealing algorithm.