Adaptive Genetic Algorithm control parameter optimization to verify the network protocol performance

J.A. Fernández Prieto, Juan R. Velasco Pérez.

Nowadays, it is important to test the computer networks under realistic traffic loads. One approach relies on integrating a Genetic Algorithm (GA) with the simulator of the system under verification. One of the main problems related to GA is to find the optimal control parameter values that it uses. Furthermore, different values may be necessary during the course of a run. Adaptive Genetic Algorithms (AGAs) have been built that dynamically adjust selected control parameters during the course of evolving a problem solution. In this paper we present a method of finding and dynamically adjusting the optimum probabilities to improve the GA performance and to drive the generation of a critical background traffic in a computer network.

PDF full paper