End-to-End QoS Guaranteed Approach Using Multi-object Genetic Algorithm in Cognitive MANETs

In this paper, we propose a novel end-to-end QoS guaranteed approach in Cognitive Mobile Ad hoc Networks (MANETs). A "black box" concept is adopted in our approach. End-to-end QoS condition is reflected by some typical QoS parameters, QoS guaranteed is implemented by the adjustments of the tunable parameters in protocol stack and the complex mechanism inside MANETs is transparent to applications and users. Multi-object genetic algorithm is used to deal with the relation between the QoS parameters and tunable parameters. QoS parameters form the fitness function and tunable parameters are encoded to the chromosome. This algorithm aims at searching the tunable parameter values which can optimize the QoS parameters. To verify the performance of genetic algorithm, its convergence is discussed in this paper. The whole approach is designed based on the cognitive method, which can enhance the adaption of our approach. In order to evaluate our approach, we design two groups of simulations based on different node mobile speed. The results show that our approach is effective and works well when topology change is rapid.