Modelling ant colony behavior in a dynamic environment and its applications to packet routing systems / Modelagem do comportamento de colônias de formigas em ambientes dinâmicos e sua aplicação em sistemas de roteamento de pacotes

AUTOR(ES)
DATA DE PUBLICAÇÃO

2009

RESUMO

The modeling of the behavior of a colony of foraging ants in a dynamic environment presented in this dissertation is based on experiments with the Argentine ant Linepithema humile under laboratory condition, using an experimental device similar to a network, consisting of tunnels and bifurcations, which is placed between the colonys nest and the food source. During the experiments, the topology of the foraging environment changes, by means of of insertion and removal of blockages in the tunnels. This dynamics requires the ants to adapt themselves in order to keep feeding the nest efficiently. In this dissertation, a newmodel for ant colony behavior is proposed, the FFAC-D (Formigas Forrageiras em Ambiente Confinado e Dinâmico is founded on evidences derived from biological studies and reproduces the experimental data, overcoming the stagnation problem observed in previous models. In the second half of this dissertation, we aim at applying the mechanisms of the FFACD Model to the packet routing problem. A review of the literature about ant based routing algorithms is presented, and, afterwards, we propose a new packet routing system based on the FFACD, the MARS (Memoryless Ants Routing System). The MARS is implemented in a discrete event simulator, the Network Simulator version 2 (NS-2), and shows a superior performance when compared to the AntNet, with respect to the convergence time and overhead (bandwidth consumption). Additionally, the MARS indicates advantages in simplicity, security and Quality of Service (QoS), over existing algorithms.

ASSUNTO(S)

aco swarm intelligence routing algorithms ciencia da computacao aco inteligência de enxames algoritmos de roteamento

ACESSO AO ARTIGO

Documentos Relacionados