非线性函数 Resources

Showing items tagged with "非线性函数"

For unknown nonlinear functions, accurately finding extremum values solely through input-output data is challenging. This problem can be solved by combining neural networks with genetic algorithms, leveraging neural networks' nonlinear fitting capabilities and genetic algorithms' nonlinear optimization abilities to locate function extrema. This article demonstrates how to optimize extremum values for nonlinear functions using neural network genetic algorithms, with implementation details including network architecture design and genetic operation parameters.

MATLAB 321 views Tagged
Download

The Extended Kalman Filter (EKF) relies solely on first-order derivatives from Taylor expansions of nonlinear functions, omitting higher-order terms. This approximation often introduces significant errors in estimating posterior state distributions, degrading both filtering algorithm performance and overall tracking system accuracy. Recently, the Unscented Kalman Filter (UKF) has emerged as an adaptive filtering alternative. Unlike EKF, UKF employs carefully designed sigma points that propagate through nonlinear functions to capture first- and second-order statistical properties of random vectors. This approach better approximates nonlinear dynamics in state equations, yielding superior estimation precision compared to EKF implementations.

MATLAB 245 views Tagged
Download

This BP Neural Network Simulink model example demonstrates a discrete transfer function implementation. The trained network can approximate arbitrary transfer functions or nonlinear functions through supervised learning algorithms.

MATLAB 210 views Tagged
Download

Wavelet Neural Network source code implementation: 1. Nonlinear function construction (nninit_test.m) - Creating custom test functions for WNN training. 2. Direct WNN approximation (Wnn_test.m) - Implements wavelet-based neural network with internal wavelet function calls. 3. Genetic Algorithm optimized WNN (GA_Wnn_test.m) - Integrates genetic algorithm components including initialization, fitness evaluation, and decoding functions for enhanced optimization.

MATLAB 214 views Tagged
Download