Ant-Inspired Approach for Resource Localization in Mobile Ad-Hoc Networks

Abstract

Introduction:

In mobile ad-hoc networks, nodes should function autonomously, and they should be able to adapt to their environment and any changes in it without any external intervention. For nodes to be able to adapt to their environment, they obviously need to have a certain knowledge about their environment and to somehow be aware of any changes within this environment. The environment of a node in a mobile ad-hoc network is made up of the nodes in the network, and how these are placed, their mobility, speed and any other characteristics that a node may enclose, such as what resources are present at which nodes in the network. One kind of environmental changes is thus changes in the network topology, which may occur frequently because of node mobility. Another possible environmental change is changes to the resource situation at one node. As we are dealing with mobile ad-hoc networks, it is important that nodes are able to monitor their environment and perform any required adaptations without any external intervention — they need to function autonomously. Autonomous, self-adapting systems are common in nature, and have been used as inspiration for solutions to a lot of computer-related problems, especially optimization problems. The most common source of inspiration is ants and their foraging behavior. A lot of research has been done on antinspired approaches to optimization problems like the routing problem, both in traditional, wired networks as well as in mobile ad-hoc networks. With this thesis, we want to look at how ants and their behavior may be used as inspiration for other kinds of problems, and to find out if such approaches may be feasible also in other scenarios than the typical optimizationproblems. As our application domain we have chosen resource localization in mobile ad-hoc networks. The purpose of the resulting solution is to enable nodes to search for any resource at any time without the need for any prior knowledge about which resources will be requested or when requests may be issued beforehand. This way, nodes may issue searches for a given resource whenever they discover a need for knowledge about the resource situation within the network.