ACO and GA metaheuristics for anomaly detection

Computer networks have become an essential technology to society, providing information and services to its users. Due to its importance, network management is necessary to maintain communication reability and security. Thus, in order to assist network administrators achieve these properties, we propose a Digital Signature of Network Segment using Flows Analysis (DSNSF), which uses the network behavior of previous weeks to predict the network traffic of a given day. For this purpose, we have developed an algorithm derived from Genetic Algorithm (GA) able to construct the DSNSF. Also, this approach is compared with a Ant Colony Optimization (ACO) modification used to the same objective. Both methods are bio-inspired models and are widely applied to optimization problems. We compare the resulting digital signature with the real traffic and use Correlation Coefficient and Normalized Square Mean Error to evaluate the performance of the algorithms.