AI Based Intrusion Detection System Using Self-Adaptive Energy Efficient BAT Algorithm for Software Defined IoT Networks
Geethu M Suresh, Minu Lalitha Madhavu · 2020
A software defined IoT network can be considered as global controller managing all the tasks of the network like resource sharing, traffic management and effective utilization of individual IoT devices connected to the network. As the network involves transmission of potential data intrusions and other vulnerabilities may arise. Traditional systems uses a standard dataset and uses evolutionary and population based algorithms for feature selection like BAT and well known classification algorithms to classify traffic as normal or attack. We propose a novel self-adaptive energy efficient BAT algorithm for achieving optimization in BAT Algorithm for Feature selection. Existing system uses swarm division for dataset clustering. But when the network traffic increases the number of features increases and non-continuous features will result. Hence swarm division doesn't works well. So a self-adaptive parallel processing strategy with energy efficiency is proposed which optimizes the existing BAT algorithm. To handle complex batch process loops within BAT algorithms, a fitness based task parallelism is implemented to get the most benefit in terms of scalability and energy efficiency.