RUTE ADAPTIF MENGGUNAKAN ALGORITMA GENETIK PADA NETWORK TELEKOMUNIKASI

• Main Author: Wibowo, Sinung
• Format: Theses
• Language: Indonesia
• Online Access: https://digilib.itb.ac.id/gdl/view/5158
• Institution: Institut Teknologi Bandung

<p>Abstract:<p align=\"justify\"> <br /> <br /> This thesis presents the result of a research on adaptive network routing table development employing genetic algorithm operators for telecommunications networks. The adaptive network routing is <p align=\"justify\"> <br /> needed to handle unpredictable network load and network reliability. A network with 12 national switching centers (Sentral Transit Nasional) that are inter-connected by 411 trunk groups has been selected as network model in this research. The adaptive routing algorithm was designed based on bulk network traffic data that was gathered every 5 minutes by a Network Traffic Management Center .<p align=\"justify\"> <br /> <br /> Genetic algorithms have been used to solve many optimization problems, including to make network routing table on call-by-call basis. Genetic operators (selection, crossover, mutation, regeneration) applied to a population in order to produce new chromosomes with better fitness. Chromosome in the population represents trunk group\'s sequence numbers of network routing table in the model network. The sequence of decimal numbers represents sequence of trunk groups that are selected to bring the traffic load from a source switch to a destination switch. Empirically, 0.9 cross over rate, 0.1 mutation rate and 10 individuals are selected as genetic algorithms parameters to solve adaptive routing problem .<p align=\"justify\"> <br /> <br /> The effectiveness of the routing algorithm was measured from the viewpoints of overall network loss. In the simulation, the model network was loaded every 5 minutes for 7 days. With static network routing table, simulation showed 5.975% network loss, but with adaptive routing table simulation showed better network loss, 0.919%. The result of the simulation showed that the adaptive routing using genetic algorithms was able to keep the network loss low even if the traffic load was increased to 117.59%.