By ever-advancing network threats existence, system security guarantee becomes critical increasingly. Generative adversarial networks (GANs) could create outcomes of striking in different fields. The ability of generation could be public when networks achieve deep comprehension based on distribution of data. Here, GANs are the promising unsupervised strategy for detecting cyber-attacks with implicitly modeling system. Selecting suitable hyperparameters is the basic problem which will influence performance of GAN. Grey Wolf optimization (GWO) is utilized for optimizing GAN hyperparameters in the thesis for improving the performance. The outcome of our test shows that our provided technique is almost efficient in order to be applied in activities of network intrusion detection also performs better than the other alike techniques of generative on dataset of NSL-KDD.
