Reconfiguration of Access Schemes in Virtual Networks of the Internet of Things by Genetic Algorithms

摘要

Virtual local area networks (VLAN) is a well-known technology of computer security in heterogeneous network infrastructures. It does not require significant computing resources. For this reason, it should find success in Internet of things. The VLAN access control scheme formation is divided into the tasks of initial configuration and reconfiguration. The paper presents an approach to the reconfiguration of VLAN access control schemes based on the improved class of genetic algorithms. Unlike the initial configuration, the reconfiguration additionally uses the previous access control scheme as input. Its search criterion is focused on minimizing the possible changes in the previous scheme. The paper shows that this problem is a special form of the Boolean matrix factorization. Main enhancements relate to generation of the initial population based on trivial solutions, using the columns of the connectivity matrix as the genes of chromosomes and applying in the fitness function the criterion of minimal cost to modify the access scheme. Experimental results demonstrate the proposed genetic algorithm has a high enough effectiveness.