Simulation of an electronic equipment control method based on an improved neural network algorithm

The neural network algorithm used to control electronic equipment is an important subject in the field of control and regulation, and its application is becoming increasingly extensive. A neural network algorithm is a very effective learning algorithm that is widely used to control electronic equipment and has the characteristics of good fault tolerance, generalisability and nonlinear mapping. However, in practice, it has many limitations (slow convergence, ease of getting stuck in local minima, forgetting old samples, etc.). To solve these problems, an advanced neural network-based algorithm is introduced, which enables the neural network to be optimised. After simulation research on the electronic equipment control method based on the improved neural network algorithm, experiments were performed, and the experimental data show that the artificial fish swarm algorithm (AFSA) training optimised PIDNN controller not only has no overshoot but also has a fast response speed and excellent dynamic performance. The optimisation by AFSA also benefits PID control, but the dynamic characteristics are slightly worse than those of the former controller by 0.1. Compared with those of the other methods evaluated, the steady-state accuracy of the optimised PIDNN is highest, with a value of 1.0. The precision values for the other two methods are 0.99 and 0.98. It can be seen from the above experimental data that training optimisation of the artificial fish swarm algorithm improves the network parameters of the PIDNN controller, thus resulting in better performance than the PIDNN controller trained by the BP algorithm.(c) 2022 Elsevier Ltd. All rights reserved.