Network partition for switched industrial Ethernet using genetic algorithm

The network partition problem in switched industrial Ethernet is analyzed, which is shown to be equivalent to a multi-objective optimization problem: the network partition should reduce the inter-network communication, and simultaneously make the network traffic be evenly distributed over the respective sub-networks. Furthermore, the switch capability must be respected when assigning devices to sub-networks, which sets constraints for the optimization problem. This is a new problem that has not been modeled before. Then genetic algorithm is proposed to search near-optimal solution for this network partition problem. When designing the fitness function and genetic operators, the communication characteristics of industrial control network, such as the existence of controller and one-way communication of field devices, are considered. Finally, a simulation research is carried out to investigate the effectiveness of the proposed genetic algorithm.