Full text
| |
| Author(s): |
Karla Vittori
Total Authors: 1
|
| Document type: | Doctoral Thesis |
| Press: | São Carlos. |
| Institution: | Universidade de São Paulo (USP). Escola de Engenharia de São Carlos (EESC/SBD) |
| Defense date: | 2005-06-27 |
| Examining board members: |
Aluízio Fausto Ribeiro Araújo;
André Carlos Ponce de Leon Ferreira de Carvalho;
Alexandre Cláudio Botazzo Delbem;
Heitor Silvério Lopes;
Leandro Nunes de Castro Silva
|
| Advisor: | Aluízio Fausto Ribeiro Araújo |
| Abstract | |
The behavior of social insects, particularly of ants, has been intensively studied in the last years, due to their capacity to perform complex tasks through interactions among simple individuals. When moving in the environment searching for food, ants deposit on the ground a chemical substance, called pheromone, to attract ants in the nest and guide them towards the source of food that was found. The process of laying/following the pheromone trails allows ants to find the shortest paths and best sources of food of the environment. With the aim to study the characteristics of the ants that contribute to their adaptation to different environment conditions, several experiments with ants have been performed and reported in this research. In this context, this thesis presents novel experiments with ants in the laboratory, in an artificial network of tunnels, where several interconnected paths lead the insects from the nest to the food source. Ants\' decision were analyzed according to the individual and collective levels, under changes in the environment, comprising the blockage/release of branches. The measure of several individual characteristics of ants allowed the development of two mathematical models of their behavior, which were applied to the non-changing (static) environment access to all branches and to changing (dynamic) access. The analysis of the collective behavior of ants in the experiments was used to compare the results derived from the simulations of the models. In general, the second proposed model was more accurate than the first one in simulating ants behavior for the static situation, however, it needs some improvements for the other situations. The satisfactory behavior of the second model led to apply its main feature, the choice taking into consideration the pheromone concentration over the branches and the ants ability to orient themselves, to solve an optimization problem, the routing in telecommunications networks. The proposed routing algorithm was evaluated under variations on the traffic level and topology of the network, and its performance was compared with two routing algorithms used by telecommunications companies, considering several performance measures. The developed algorithm produced encouraging results, suggesting the possibility to apply the framework of the proposed model to other optimization problems. (AU) | |