Dynamic RMSA in elastic optical networks with an adaptive genetic algorithm

We develop an adaptive and efficient genetic algorithm (GA) to solve the dynamic routing, modulation and spectrum assignments (RMSA) for elastic O-OFDM networks. The algorithm offers an efficient way of serving the dynamic lightpath requests based on the current network status at each service provision time. The GA is designed for multi-objective optimization. For low traffic cases when there is no blocking, the GA minimizes the maximum number of slots required on any fiber in the network; otherwise, it minimizes the blocking probability. The performance of the proposed GA is evaluated in dynamic RMSA simulations with the 14-node NSFNET and the 28-node US Backbone topologies, and the results show that it converges within 25 generations. The simulation results also verify that the GA-RMSA outperforms several existing algorithms by providing more load-balanced network provisioning solutions with lower blocking probabilities. Specifically, when the traffic load is same, the GA can achieve more than one order-of-magnitude reduction on blocking probability. To the best of our knowledge, this is the first attempt to solve dynamic RMSA in elastic O-OFDM networks with a GA.