Using Improved Genetic Algorithms to Design Switched Industrial Ethernet Topology
Bingwen Wang · Computer Engineering and Science · 2007
This paper proposes an approach which uses improved genetic algorithms to optimize and design industrial control networks. With the redundant topology structure and the communication characteristics of switched industrial Ethernet, the problem of network partitioning is analyzed to be equivalent to a multi-objective optimization problem: topology optimization should reduce the inter-network communications, and simultaneously allow the network traffic to be evenly distributed over all sub-networks. Moreover, the switch’s capability must be respected, which sets constraints for the optimization problem. In the improved genetic algorithms,the coding strategy,the probability of crossover and the probability of mutation are improved, which conquers the shortcomings of premature convergence in simple genetic algorithms. The simulation result shows that the algorithm is effective.