Adaptive Inverse Induction Machine Control Based on Variable Learning Rate BP Algorithm

Abstract

The adaptive inverse control technology is utilized for induction machine (IM) control. Adaptive inverse control is actually an open-loop control scheme and so in the adaptive inverse control the instability problem caused by feedback control is avoided and the better dynamic performances can also be achieved. Linear LMS technique of adaptive inverse control is extended to control the MIMO, nonlinear IM based on BP neural network. And the BP algorithm is improved by using variable learning rate. Simulation study is made to validate the effectiveness of the control scheme.Index Terms -adaptive inverse control; induction machine; neural network; variable learning rate

Gradient Norm Aware Minimization Seeks First-Order Flatness and Improves Generalization

“…Some studies [20,69] have demonstrated that current optimization approaches, such as SGD [55], Adam [40], AdamW [51] and others [19,48] affect generalization. Some previous literature finds that Adam is more vulnerable to sharp minima than SGD [65], which results in worse generalization ability [22,28,68].…”

“…Some previous literature finds that Adam is more vulnerable to sharp minima than SGD [65], which results in worse generalization ability [22,28,68]. Some following works [10,52,69,76] propose generalizable optimizers to address this problem. However, it can be a trade-off between generalization ability and convergence speed [19,38,48,69,76].…”

A High-Accuracy and Fast Retrieval Method of Atmospheric Parameters Based on Genetic-BP

“…The learning rate variable BP algorithm [37,41] has an adaptive ability of adjusting the learning rate gradient descent in terms of the variation of error. The learning rate will increase if the error decreases; otherwise, the adjustment is wrong, and the step size should be reduced.…”

“…However, the network convergence speed of the BP neural network algorithm is a little slow due to the problem of finding the first derivative of the objective function. In general, for improving the convergence speed, there exists a kind of improved heuristic algorithms, such as the momentum Back Propagation [35][36] and variable learning rate Back Propagation [37]. And also, there exists another kind of improved numerical optimization method, such as the conjugate gradient Back Propagation [38] and the Levenberg-Marquardt Back Propagation (LMBP) [39][40].…”

Intelligent coordinated controller design for a 600MW supercritical boiler unit based on expanded-structure neural network inverse models

“…It has drawn much Contents lists available at ScienceDirect journal homepage: www.elsevier.com/locate/conengprac attention in engineering applications with its advantages of clear physical concept, being intuitive and easy to understand (Yuan & Guo, 1994;Han et al, 2001;Dai, He, Zhang, & Zhang, 2003;Dai, 2005;Xie, Zhang, & Xiao, 2007). But solving for the inverse system model of a complex multivariable system is a bottleneck.…”