基于自适应遗传算法优化BP神经网络

Optimized BP Neural Network Based on Self-adapted Genetic Algorithm

传统的BP算法具有简单可塑的优点,但是存在着容易陷入局部极值、收敛速度慢等无法克服的缺陷。遗传算法是一种全局搜索算法,具有很强的全局搜索能力。因此,文中设计了一种自适应遗传算法优化BP神经网络的方法,并通过对数字信号的实例仿真,证明了经过改进优化后的BP神经网络具有很好的应用前景。

Conventional BP algorithm has merit of plasticity, but presents insuperable bug that it easily gets into partial extremum, and has slow speed of convergence. Genetic algorithm is an wholly-searching algo-rithm, which has strong capability of wholly-search. Consequently, the paper design a BP neural network algo- rithm optimized by self-adapted genetic algorithm. It proves that optimized BP neural network has very excellent applied foreground through the simulation of digital information.