A metamodel-based optimization method for mobile ad hoc networks

To address the dilemma between the accuracy of simulation-model and runtime efficiency which take place in the simulation-based optimization of mobile ad hoc networks, the metamodel-based optimization methodology is investigated to optimize the performance of networks. Hybrid optimization strategy is proposed in detail, which integrated Kriging metamodel, evolution algorithm and trust region algorithm. With this optimization strategy, a nonlinear parameter settings optimization problem is resolved. The results show, the proposed metamodel-based optimization strategy can improve the efficiency of computation-intensive process, it can be successfully applied to the domain of complex systems such as mobile ad hoc networks, facilitating system modeling, analysis and optimization.