Genetic algorithm for optimizing and arrangement of queue in virtual networks

Network visualization provides methods that simplify resource management, deal with connectivity and bandwidth constraints, and enable virtual network-connected machines through either a common layer 2 or layer 3 IP network of TCP/IP protocol suite. This paper presents an alternative optimization solution in maintaining the packets queue on a virtual router by using Genetic Algorithm. The objective of this research is to reduce a cost of network resources such as memory and time processes on a router. It is important for a company with limited cost when implements a network infrastructure. We defined a crossover probability and a mutation probability to 0.90 and 0.05. In our experiments, we still have unsatisfied results due to the average and the standard deviation of fitness values and slots needed in queues are 58,238.10 & 139,575.45 and 106.20 & 82.51, respectively. These values should be better by defining an appropriate fitness function in our next experiments. We still continue the research by examining an appropriate fitness function, Genetic Algorithm as a classifier, dynamic slot size, and also comparing with current queue managements such as first-in first-out and weighted fair queue.