基于智能优化算法的神经网络安全设计与优化

Design and Optimization of Neural Network Security Based on Intelligent Optimization Algorithm

为了解决网络安全评估模型收敛速度慢、评估指标不完善、模型参数不完备导致的网络安全预测不准确的问题,设计了基于智能优化算法的神经网络状态评估和预测优化模型.首先,基于智能优化算法的并行性、高效性对神经网络评估进行优化,并划分相应的安全等级.其次,利用蚁群算法获取神经网络状态预测模型源码值并对其进行优化,使神经网络状态更加高效和稳定.每个模型都设计实验验证遗传算法对神经网络模型的优化和提升.

In order to solve problems of inaccurate network security prediction caused by slow convergence speed, imperfect evaluation index and imperfect model parameters, a neural network state evaluation and prediction optimization model based on intelligent optimization algorithm is designed. Firstly, based on the parallelism and efficiency of intelligent optimization algorithm, the evaluation of neural network is optimized, and the corresponding security level is divided. Secondly, the ant colony algorithm is used to obtain the source code value of the neural network state prediction model and optimize it to make the neural network state more efficient and stable. Experiment is designed for each model to verify the optimization and improvement of the neural network model by genetic algorithm.