Network Partition of Switched Industrial Ethernet by Using Novel Particle Swarm Optimization

Abstract

In this paper, a novel particle swarm optimization is proposed to solve integer programming problem, which is name as an adaptive inertia weight integer programming particle swarm optimization (AIWIP-PSO). Then it is applied to solve the industrial Ethernet network partition problem, after the problem is formulated as multi-objective integer programming problem. This multi-objective integer programming problem includes two objectives: one is the minimum of the inter-network communication; the other is the evenness of the network traffic between the sub-networks. Furthermore, the switch capability must be considered when assigning devices to sub-networks. Two network partition cases are chosen for simulations. The results obtained by using AIWIP-PSO are compared with those from genetic algorithm. Simulation results show that the AIWIP-PSO is very effective and outperforms genetic algorithm in terms of the performance index